fix(edu-master): bound session refreshes and expire stale cookies

ClusterIP access to api@internal from Homarr pod. LAN-reachable on 192.168.80.2:8080, accepted.

.gitea/workflows/ci.yaml

@@ -90,11 +90,8 @@ jobs:
           set -euo pipefail
           tools_dir="$(bash .gitea/workflows/install-ci-tools.sh shellcheck)"
           export PATH="$tools_dir:$PATH"
-          # userbot/ is a git subtree synced from forust/userbot, so its shell
-          # scripts are upstream's to maintain, not ours. Linting them would let a
-          # routine subtree pull turn the deploy gate red on code we do not own.
           mapfile -t scripts < <(
-            git ls-files '*.sh' ':(glob)**/*.bash' ':!userbot/**'
+            git ls-files '*.sh' ':(glob)**/*.bash'
           )
           if [ "${#scripts[@]}" -eq 0 ]; then
             echo "No shell scripts found."
@@ -144,6 +141,7 @@ jobs:
           export PATH="$tools_dir:$PATH"
           ruff check .
           ruff format --check .
+          python3 -m unittest discover -s tests -v
 
   lint-yaml:
     runs-on: [self-hosted, linux, arch, homelab]
@@ -197,155 +195,6 @@ jobs:
 
           hadolint -c .hadolint.yaml "${dockerfiles[@]}"
 
-  # Known, accepted, and recorded. Each line is a real advisory against a
-  # package we build into the panel image, kept in this workflow rather than in
-  # the package manifest so that a subtree sync from forust/userbot cannot
-  # silently widen the exemption.
-  #
-  # starlette is the reason this job is not simply "fail on everything":
-  # fastapi 0.115.12 pins `starlette<0.47.0`, and the fixes for the last four
-  # below need 0.49.1 through 1.3.1, so clearing them means a jump from fastapi
-  # 0.115.12 to 0.141.x. That is upstream's call, not a drive-by in a lint
-  # commit. Of the seven, four are reachable here in principle: 1942 is a
-  # crafted Range header hitting FileResponse, and the panel serves its built
-  # SPA through exactly that; 249 is request.form() ignoring max_fields for
-  # x-www-form-urlencoded, which is the login form; 1941 is a large multipart
-  # body blocking the event loop; 161 and 248 are unvalidated Host and request
-  # path reaching request.url. 2280 needs HTTPEndpoint, which the panel does
-  # not use, and 2281 is Windows-only, and this deploys on Linux.
-  #
-  # The panel answers on userbot.workstation.internal and has no public
-  # forust.xyz route, which is what keeps the four reachable ones from being
-  # an internet-facing DoS. It still manages Telegram credentials.
-  #
-  # Deleting an entry here is how you accept a new advisory, so the diff says
-  # so out loud.
-  scan-deps:
-    runs-on: [self-hosted, linux, arch, homelab]
-    timeout-minutes: 15
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4
-
-      - name: Audit the Python dependencies that ship in the image
-        shell: bash
-        run: |
-          set -euo pipefail
-          tools_dir="$(bash .gitea/workflows/install-ci-tools.sh pip-audit)"
-          export PATH="$tools_dir:$PATH"
-          # requirements.txt, not requirements-dev.txt: this is what the image
-          # installs, and the test tooling is not a shipped attack surface.
-          pip-audit -r userbot/panel/backend/requirements.txt --strict \
-            --ignore-vuln CVE-2025-67720 \
-            --ignore-vuln PYSEC-2026-161 \
-            --ignore-vuln PYSEC-2026-1941 \
-            --ignore-vuln PYSEC-2026-1942 \
-            --ignore-vuln PYSEC-2026-2280 \
-            --ignore-vuln PYSEC-2026-2281 \
-            --ignore-vuln PYSEC-2026-248 \
-            --ignore-vuln PYSEC-2026-249
-
-      # devDependencies are excluded on purpose. `npm audit` on the full tree
-      # reports 7 findings, and every one of them is a build- or test-time
-      # package: the esbuild CORS advisory needs a vite dev server serving to
-      # the internet, and nanoid's infinite loop needs a custom generator
-      # called with size 0, which postcss does not do. None of them are in the
-      # 91 kB bundle the panel serves. The one production finding, devalue
-      # via svelte, is moderate, which is where --audit-level draws the line;
-      # this fails on the next high or critical one.
-      - name: Audit the production npm dependencies
-        shell: bash
-        run: |
-          set -euo pipefail
-          # The pinned node, not whatever the runner has. Its system node is a
-          # rolling Arch package: during this very push its npm was missing
-          # entirely, and an hour later it was npm 12 on node 26. Both are the
-          # wrong major anyway — the panel image is node:22-alpine.
-          tools_dir="$(bash .gitea/workflows/install-ci-tools.sh node)"
-          export PATH="$tools_dir:$PATH"
-          cd userbot/panel/frontend
-          npm ci
-          npm audit --omit=dev --audit-level=high
-
-  test-backend:
-    runs-on: [self-hosted, linux, arch, homelab]
-    timeout-minutes: 15
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4
-
-      # 25 tests over the panel's pydantic models, its auth flow, the SPA
-      # fallback and the Kubernetes client it shells out with. They existed and
-      # had never been executed by anything.
-      #
-      # Note that userbot/ is a subtree synced from forust/userbot, so a routine
-      # sync can turn this red on upstream's code. Unlike the shellcheck job,
-      # which skips that tree because style disagreements there are ours to
-      # lose, a failing test here is a real defect in a service we deploy.
-      - name: Run the panel backend test suite
-        shell: bash
-        run: |
-          set -euo pipefail
-          tools_dir="$(bash .gitea/workflows/install-ci-tools.sh uv)"
-          export PATH="$tools_dir:$PATH"
-
-          # A venv in a temp dir rather than a checked-out one: the runner is
-          # shared, and a leftover .venv would let a dependency the
-          # requirements no longer pin still satisfy an import.
-          #
-          # --python is not optional. uv otherwise takes whatever interpreter it
-          # finds first, and which one that is depends on the machine: this
-          # runner runs jobs on the host, where the only interpreter is 3.14,
-          # and pyrogram's sync.py calls the bare asyncio.get_event_loop() that
-          # 3.14 no longer auto-creates, so three tests fail at collection. The
-          # image is python:3.13-slim, so 3.13 is also the version worth
-          # testing: uv fetches a managed build of it when the host has none,
-          # which is what makes this job independent of the runner.
-          venv="$(mktemp -d)/venv"
-          uv venv --python 3.13 --quiet "$venv"
-          uv pip install --quiet --python "$venv/bin/python" \
-            -r userbot/panel/backend/requirements-dev.txt
-
-          # `python -m`, not bare `pytest`: the tests import `app.*` relative to
-          # the backend directory, which only works if the cwd is on sys.path,
-          # and only `python -m` puts it there.
-          cd userbot/panel/backend
-          "$venv/bin/python" -m pytest tests/ -q
-
-  test-frontend:
-    runs-on: [self-hosted, linux, arch, homelab]
-    timeout-minutes: 15
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4
-
-      # One `npm ci` for both checks below: it is by far the slowest part of
-      # this job, and a second one would learn nothing the first did not.
-      #
-      # `npm ci`, not `npm install`, for the same reason the Dockerfile uses it:
-      # the lockfile is what makes the tree that gets checked the tree that
-      # gets shipped.
-      - name: Type-check and test the panel frontend
-        shell: bash
-        run: |
-          set -euo pipefail
-          # The pinned node, not whatever the runner has. Its system node is a
-          # rolling Arch package: during this very push its npm was missing
-          # entirely, and an hour later it was npm 12 on node 26. Both are the
-          # wrong major anyway — the panel image is node:22-alpine.
-          tools_dir="$(bash .gitea/workflows/install-ci-tools.sh node)"
-          export PATH="$tools_dir:$PATH"
-          cd userbot/panel/frontend
-          npm ci
-
-          # svelte-check has been a devDependency all along with no script
-          # pointing at it, so the type errors it reports had nowhere to
-          # surface. It is clean today, which is the only reason it can be a
-          # gate: it stops at whatever upstream introduces rather than
-          # reporting a backlog we inherited.
-          npm run check
-          npm test
-
   validate:
     runs-on: [self-hosted, linux, arch, homelab]
     timeout-minutes: 20
@@ -464,8 +313,12 @@ jobs:
 
   build:
     needs:
+      # The panel's scan-deps/test-backend/test-frontend jobs gated here until
+      # userbot moved to its own repo; upstream's code is upstream's gate now.
+      # The rule is unchanged: publishing and passing the checks are the same
+      # gate, so a commit that fails any of these still cannot move :prod.
       [lint-actionlint, lint-shellcheck, lint-compose, lint-prettier, lint-ruff, lint-yaml, lint-dockerfiles, validate]
-    if: github.event_name != 'pull_request' && (github.ref_name == 'main' || github.ref_name == 'dev')
+    if: github.event_name != 'pull_request' && (github.ref_name == 'main' || github.ref_name == 'dev') && !startsWith(github.ref_name, 'renovate/')
     runs-on: [self-hosted, linux, arch, homelab]
     timeout-minutes: 60
     outputs:
@@ -480,12 +333,20 @@ jobs:
         id: services
         shell: bash
         run: |
+          set -euo pipefail
           base="${{ github.event.before }}"
           if [ -z "$base" ] || [ "$base" = "0000000000000000000000000000000000000000" ]; then
             base="$(git rev-list --max-parents=0 HEAD)"
           fi
 
-          mapfile -t changed_files < <(git diff --name-only "$base" "${GITHUB_SHA}")
+          # A failed diff used to leave changed_files empty, which reads exactly
+          # like "nothing to build": the job went green having built nothing and
+          # the tag never moved. The status is checked, not assumed.
+          if ! changed="$(git diff --name-only "$base" "${GITHUB_SHA}")"; then
+            echo "::error::cannot diff ${base}..${GITHUB_SHA}"
+            exit 1
+          fi
+          mapfile -t changed_files <<<"$changed"
 
           services=()
 
@@ -505,15 +366,9 @@ jobs:
 
           for file in "${changed_files[@]}"; do
             case "$file" in
-              dtek_notif/*)
-                add_service dtek_notif
-                ;;
               errorpages/*)
                 add_service errorpages
                 ;;
-              userbot/*)
-                add_service userbot
-                ;;
               homepages/*)
                 add_service homepages
                 ;;
@@ -533,55 +388,63 @@ jobs:
           echo "services=$(paste -sd, /tmp/services.txt)" >> "$GITHUB_OUTPUT"
 
       - name: Log in to registry
-        if: steps.services.outputs.services != ''
+        # The pin step below also writes (manifest PUTs), and it runs on every
+        # main push — including manifest-only ones where services is empty. A
+        # stale persistent login on the old runner used to mask this; a clean
+        # runner pushes anonymously and gets 401.
+        if: steps.services.outputs.services != '' || github.ref_name == 'main'
         shell: bash
+        # Through env, not by substitution into the script. A secret written
+        # into a run: block is pasted into the shell source before bash parses
+        # it, so a password containing a quote, a backtick or $(...) becomes
+        # code that runs. Masking the value in the log does not prevent that.
+        env:
+          REGISTRY_USERNAME: ${{ secrets.REGISTRY_USERNAME }}
+          REGISTRY_PASSWORD: ${{ secrets.REGISTRY_PASSWORD }}
         run: |
-          echo "${{ secrets.REGISTRY_PASSWORD }}" | docker login "${REGISTRY}" \
-            -u "${{ secrets.REGISTRY_USERNAME }}" \
+          set -euo pipefail
+          printf '%s' "$REGISTRY_PASSWORD" | docker login "${REGISTRY}" \
+            -u "$REGISTRY_USERNAME" \
             --password-stdin
 
       - name: Build and push changed images
         if: steps.services.outputs.services != ''
         shell: bash
         run: |
+          # This step was the one run: block in the workflow without it, and it
+          # is the one that cannot afford it: a docker push that failed partway
+          # through the loop used to be followed by more pushes, the loop's exit
+          # status came from the last one, and the job went green with half the
+          # images missing from the registry.
+          set -euo pipefail
           IFS=, read -r -a services <<< "${{ steps.services.outputs.services }}"
 
+          # Tags for this push. The commit-pinned name is the point of this
+          # step: the deploy resolves it in preference to :prod, so a deploy
+          # that sat in the queue behind a later push still gets the build of
+          # the commit CI validated, instead of whatever :prod points at by the
+          # time it runs. See render_pinned in deploy-lib.sh.
+          commit_tag=""
+          if [ "${GITHUB_REF_NAME}" = "main" ]; then
+            commit_tag="sha-${GITHUB_SHA:0:12}"
+          fi
+
+          set_tags() {
+            tags=()
+            case "${GITHUB_REF_NAME}" in
+              main) tags+=("main" "prod") ;;
+              dev) tags+=("dev") ;;
+            esac
+            if [ -n "$commit_tag" ]; then
+              tags+=("$commit_tag")
+            fi
+          }
+
           for service in "${services[@]}"; do
             case "$service" in
-              dtek_notif)
-                image="${REGISTRY}/forust/dtek-notif"
-                tags=()
-                case "${GITHUB_REF_NAME}" in
-                  main)
-                    tags+=("main" "prod")
-                    ;;
-                  dev)
-                    tags+=("dev")
-                    ;;
-                esac
-                build_args=()
-                for tag in "${tags[@]}"; do
-                  build_args+=(-t "${image}:${tag}")
-                done
-                docker build \
-                  --cache-from "type=registry,ref=${image}:buildcache" \
-                  --cache-to "type=registry,ref=${image}:buildcache,mode=max" \
-                  "${build_args[@]}" dtek_notif
-                for tag in "${tags[@]}"; do
-                  docker push "${image}:${tag}"
-                done
-                ;;
               errorpages)
                 image="${REGISTRY}/forust/error-pages"
-                tags=()
-                case "${GITHUB_REF_NAME}" in
-                  main)
-                    tags+=("main" "prod")
-                    ;;
-                  dev)
-                    tags+=("dev")
-                    ;;
-                esac
+                set_tags
                 build_args=()
                 for tag in "${tags[@]}"; do
                   build_args+=(-t "${image}:${tag}")
@@ -594,40 +457,6 @@ jobs:
                   docker push "${image}:${tag}"
                 done
                 ;;
-              userbot)
-                tags=()
-                case "${GITHUB_REF_NAME}" in
-                  main)
-                    tags+=("main" "prod")
-                    ;;
-                  dev)
-                    tags+=("dev")
-                    ;;
-                esac
-                for target in runtime panel; do
-                  case "$target" in
-                    runtime)
-                      context="userbot"
-                      image="${REGISTRY}/forust/userbot"
-                      ;;
-                    panel)
-                      context="userbot/panel"
-                      image="${REGISTRY}/forust/userbot-panel"
-                      ;;
-                  esac
-                  build_args=()
-                  for tag in "${tags[@]}"; do
-                    build_args+=(-t "${image}:${tag}")
-                  done
-                  docker build \
-                    --cache-from "type=registry,ref=${image}:buildcache" \
-                    --cache-to "type=registry,ref=${image}:buildcache,mode=max" \
-                    "${build_args[@]}" "$context"
-                  for tag in "${tags[@]}"; do
-                    docker push "${image}:${tag}"
-                  done
-                done
-                ;;
               homepages)
                 for variant in forust xdfnx; do
                   case "$variant" in
@@ -638,15 +467,7 @@ jobs:
                       image="${REGISTRY}/forust/xdfnx-homepage"
                       ;;
                   esac
-                  tags=()
-                  case "${GITHUB_REF_NAME}" in
-                    main)
-                      tags+=("main" "prod")
-                      ;;
-                    dev)
-                      tags+=("dev")
-                      ;;
-                  esac
+                  set_tags
                   build_args=()
                   for tag in "${tags[@]}"; do
                     build_args+=(-t "${image}:${tag}")
@@ -672,15 +493,7 @@ jobs:
                       image="${REGISTRY}/forust/webinar-checker"
                       ;;
                   esac
-                  tags=()
-                  case "${GITHUB_REF_NAME}" in
-                    main)
-                      tags+=("main" "prod")
-                      ;;
-                    dev)
-                      tags+=("dev")
-                      ;;
-                  esac
+                  set_tags
                   build_args=()
                   for tag in "${tags[@]}"; do
                     build_args+=(-t "${image}:${tag}")
@@ -696,3 +509,42 @@ jobs:
                 ;;
             esac
           done
+
+      # Every image the tree names has to carry the commit-pinned name, not only
+      # the ones this push rebuilt. A push that touches nothing but manifests
+      # builds nothing, and its deploy would then find no commit-pinned tag to
+      # resolve and quietly fall back to the moving :prod - which is the whole
+      # failure the commit-pinned name exists to remove.
+      #
+      # Re-tagging copies the manifest list and transfers no layers, so pinning
+      # six images that already exist costs six registry writes.
+      #
+      # The list is derived from the tree rather than written out here, so an
+      # image added to a manifest is covered without a second place to update.
+      - name: Pin the commit name on the images this push did not rebuild
+        if: github.ref_name == 'main'
+        shell: bash
+        run: |
+          set -euo pipefail
+          commit_tag="sha-${GITHUB_SHA:0:12}"
+          mapfile -t repos < <(
+            git grep -hoE 'gcr\.forust\.xyz/forust/[A-Za-z0-9._-]+' -- '*.yaml' '*.yml' \
+              | sort -u
+          )
+          if [ "${#repos[@]}" -eq 0 ]; then
+            echo "No own images referenced by the tree."
+            exit 0
+          fi
+          echo "pinning ${#repos[@]} image(s) to $commit_tag"
+          for repo in "${repos[@]}"; do
+            if docker buildx imagetools inspect "$repo:$commit_tag" >/dev/null 2>&1; then
+              echo "  already built by this push: ${repo##*/}"
+              continue
+            fi
+            if ! docker buildx imagetools inspect "$repo:prod" >/dev/null 2>&1; then
+              echo "  WARNING: ${repo##*/} has no :prod to pin and no build produced it"
+              continue
+            fi
+            docker buildx imagetools create --tag "$repo:$commit_tag" "$repo:prod"
+            echo "  pinned ${repo##*/}"
+          done

.gitea/workflows/deploy-lib.sh

@@ -233,20 +233,73 @@ registry_digest() {
   # pipefail reports the rightmost non-zero stage, so a ref the registry does not
   # have would abort the caller at the assignment instead of yielding an empty
   # string. The callers check for empty themselves and report it by name.
-  docker manifest inspect "$1" 2>/dev/null \
-    | jq -r --arg arch "$arch" '
-        .manifests[]?
-        | select(.platform.os == "linux" and .platform.architecture == $arch)
-        | .digest
-      ' 2>/dev/null \
-    | head -1 || true
+  #
+  # Retried with a hard timeout because the registry has a known hang mode (and
+  # a known blink mode: a single failed lookup aborts the whole apply file in
+  # render_pinned). A short sleep between attempts lets a restarting registry
+  # come back instead of failing the deploy on one bad second.
+  local attempt=0 digest=""
+  while [ "$attempt" -lt 3 ]; do
+    digest="$(timeout 25s docker manifest inspect "$1" 2>/dev/null \
+      | jq -r --arg arch "$arch" '
+          .manifests[]?
+          | select(.platform.os == "linux" and .platform.architecture == $arch)
+          | .digest
+        ' 2>/dev/null \
+      | head -1 || true)"
+    [ -n "$digest" ] && break
+    attempt=$((attempt + 1))
+    if [ "$attempt" -lt 3 ]; then
+      echo "WARNING: registry lookup of $1 failed (attempt $attempt/3), retrying in 5s" >&2
+      sleep 5
+    fi
+  done
+  printf '%s' "$digest"
+}
+
+# The commit this deploy is for: what CI validated, or - on a manual dispatch,
+# whatever stage_preflight just checked out.
+deploy_commit() {
+  local c="${DEPLOY_SHA:-}"
+  [ -n "$c" ] || c="$(git -C "$REPO" rev-parse HEAD 2>/dev/null || true)"
+  printf '%.12s' "${c:-}"
+}
+
+# Resolves one of our image refs to the digest THIS commit's build produced.
+#
+# A manifest naming `:prod` names a pointer, not a version, and the deploy
+# resolves it when the apply runs - which is not when CI ran it. Deploy runs are
+# queued rather than cancelled (see deploy.yaml), so two pushes in a row leave
+# the first deploy resolving the second push's build: the right manifests with
+# the wrong code, and nothing anywhere reports it. ci therefore publishes every
+# image it ships under `sha-<commit12>`, a name that cannot move, and that is
+# the name resolved here.
+#
+# The fallback to the plain tag is for an image this pipeline never built. It
+# reports itself, because a fallback nobody sees is the failure this removes.
+pinned_digest() {
+  local ref="$1" commit pinned
+  commit="$(deploy_commit)"
+  if [ -n "$commit" ]; then
+    pinned="$(registry_digest "${ref%:*}:sha-$commit")"
+    if [ -n "$pinned" ]; then
+      printf '%s' "$pinned"
+      return 0
+    fi
+  fi
+  pinned="$(registry_digest "$ref")"
+  if [ -n "$pinned" ]; then
+    echo "WARNING: ${ref} carries no sha-${commit:-<unknown>} tag; resolved the moving tag instead" >&2
+  fi
+  printf '%s' "$pinned"
 }
 
 # Rewrites our own images to immutable digests on the way into the cluster.
 # Reads a manifest stream on stdin, writes the pinned stream to stdout.
 #
-# A digest is not knowable when a manifest is written, so it is resolved here, at
-# apply time, and never committed: git keeps a readable `:prod` tag. That is what
+# A digest is not knowable when a manifest is written, so it is never committed:
+# git keeps a readable `:prod` tag and the exact bytes are chosen here, at apply
+# time, from the tag ci published for the commit being deployed. That is what
 # makes rollback mean something. `kubectl rollout undo` restores the previous
 # ReplicaSet's pod template verbatim, and a template naming a digest restores the
 # exact bytes that were serving before. A template naming a moving tag does not —
@@ -271,7 +324,7 @@ render_pinned() {
 
   while read -r ref; do
     [ -n "$ref" ] || continue
-    digest="$(registry_digest "$ref")"
+    digest="$(pinned_digest "$ref")"
     if [ -z "$digest" ]; then
       echo "ERROR: cannot resolve ${ref} in the registry; applying nothing." >&2
       echo "       The build job has to push that tag before the deploy resolves it." >&2
@@ -339,7 +392,7 @@ restart_stale_images() {
   while read -r ns target image; do
     [ -n "${target:-}" ] || continue
     if [ -z "${digests[$image]:-}" ]; then
-      digests[$image]="$(registry_digest "$image")"
+      digests[$image]="$(pinned_digest "$image")"
     fi
     want="${digests[$image]}"
     if [ -z "$want" ]; then
@@ -451,6 +504,14 @@ rollback_workloads() {
   local -a recovered=()
   while read -r kind ns name; do
     [ -n "${kind:-}" ] || continue
+    # Helm-owned workloads are already rolled back by the release's --rollback-on-failure
+    # upgrade. `rollout undo` here would step back to the revision Helm just
+    # escaped (the failed one), so leave them for the operator instead.
+    if kubectl get "${kind}/${name}" -n "$ns" -o jsonpath='{.metadata.annotations}' 2>/dev/null | grep -q 'meta.helm.sh/release-name'; then
+      echo "  skip (helm-managed, needs manual check): ${kind}/${ns}/${name}"
+      unrecovered+=("${kind}/${ns}/${name} (helm-managed)")
+      continue
+    fi
     if kubectl rollout undo "${kind}/${name}" -n "$ns" >/dev/null 2>&1 \
       && kubectl rollout status "${kind}/${name}" -n "$ns" --timeout="${ROLLOUT_TIMEOUT}s" >/dev/null 2>&1; then
       echo "  rolled back: ${kind}/${ns}/${name}"
@@ -489,6 +550,66 @@ helm_repo_for() {
   esac
 }
 
+# helm_release_status <release> <namespace>
+# Prints the release status in lowercase (deployed, failed, pending-rollback,
+# ...) or "not-found" when the release does not exist yet.
+helm_release_status() {
+  local out
+  if ! out="$(helm status "$1" -n "$2" 2>&1)"; then
+    echo "not-found"
+    return 0
+  fi
+  awk '/^STATUS:/{print $2}' <<<"$out" | tr '[:upper:]' '[:lower:]'
+}
+
+# recover_pending_release <release> <namespace>
+# Rolls a release out of a pending-* state left by a failed upgrade with --rollback-on-failure
+# whose own rollback never completed. Without this every future upgrade errors
+# out until a human runs `helm rollback`. Passes through releases that are not
+# pending (deployed, failed, not-found). Returns non-zero when the release is
+# still not recoverable, so the pipeline fails loud instead of wedging.
+recover_pending_release() {
+  local release="$1" namespace="$2" status
+  status="$(helm_release_status "$release" "$namespace")"
+  case "$status" in
+    pending-upgrade|pending-rollback|pending-install)
+      log "Release $release is $status, rolling back to the last deployed revision"
+      if ! helm rollback "$release" -n "$namespace" --wait --timeout 10m >/dev/null 2>&1; then
+        echo "WARN: helm rollback of $release did not complete"
+        return 1
+      fi
+      status="$(helm_release_status "$release" "$namespace")"
+      if [ "$status" != "deployed" ]; then
+        echo "WARN: $release is $status after rollback"
+        return 1
+      fi
+      ;;
+  esac
+  return 0
+}
+
+# wait_for_calm <stage>
+# The deploy itself is heavy enough to melt this single node (helm churn plus
+# apply churn drove load past 40, killed netbird/ssh, left helm pending-*).
+# Never pile a heavy step onto an already-hot node: wait up to 10 minutes for
+# the 1-minute load average to drop below the ceiling, then proceed anyway
+# with a warning so a permanently busy node cannot wedge the pipeline forever.
+wait_for_calm() {
+  local load waited=0
+  while [ "$waited" -lt 600 ]; do
+    load="$(cut -d' ' -f1 /proc/loadavg | cut -d. -f1)"
+    if [ "$load" -lt 28 ]; then
+      return 0
+    fi
+    if [ "$((waited % 60))" -eq 0 ]; then
+      log "$1: load $load, waiting for calm (<28)..."
+    fi
+    sleep 15
+    waited=$((waited + 15))
+  done
+  echo "WARN: $1: node still loaded ($load) after 10m, proceeding anyway"
+}
+
 upgrade_helm_releases() {
   local entry release chart namespace version values marker repo
   for entry in ${HELM_RELEASES[@]+"${HELM_RELEASES[@]}"}; do
@@ -509,13 +630,33 @@ upgrade_helm_releases() {
     helm repo add "${repo%% *}" "${repo#* }" >/dev/null 2>&1 || true
     helm repo update "${repo%% *}" >/dev/null 2>&1 || true
     log "Upgrading $release ($chart $version)"
-    # --atomic rolls the release back when the upgrade times out or the workloads
-    # it touches never become ready, so a bad chart bump is not left half applied.
-    helm upgrade --install "$release" "$chart" \
+    wait_for_calm "helm $release"
+    # A previous run with --rollback-on-failure whose own rollback never finished leaves the
+    # release in pending-*, which blocks every future upgrade. Recover first
+    # so one wedged revision cannot wedge the pipeline forever.
+    if ! recover_pending_release "$release" "$namespace"; then
+      echo "ERROR: $release is stuck and automatic rollback did not recover it, run 'helm rollback $release -n $namespace' by hand."
+      return 1
+    fi
+    # --rollback-on-failure (+ --wait) rolls the release back when the upgrade
+    # times out or the workloads it touches never become ready, so a bad chart
+    # bump is not left half applied. (--atomic was this combo; deprecated.)
+    if ! helm upgrade --install "$release" "$chart" \
       --namespace "$namespace" \
       --version "$version" \
       --values "$REPO/$values" \
-      --atomic --cleanup-on-fail --timeout 10m
+      --wait --rollback-on-failure --cleanup-on-fail --timeout 10m; then
+      echo "WARN: upgrade of $release failed, checking release state"
+      # --rollback-on-failure already attempted its own rollback; finish the job when that
+      # rollback never completed, otherwise the release stays pending-* and
+      # blocks every future run.
+      if ! recover_pending_release "$release" "$namespace"; then
+        echo "ERROR: upgrade of $release failed and the release did not recover, run 'helm rollback $release -n $namespace' by hand."
+      else
+        echo "ERROR: upgrade of $release failed (release is back on its previous revision)."
+      fi
+      return 1
+    fi
   done
 }
 
@@ -645,6 +786,7 @@ stage_apply_k8s() {
     fi
   fi
   upgrade_helm_releases
+  wait_for_calm "apply resources"
   if [ "${#other_files[@]}" -gt 0 ]; then
     log "Applying resources (${#other_files[@]} files, our images pinned to digests)"
     for m in "${other_files[@]}"; do
@@ -661,19 +803,6 @@ stage_apply_k8s() {
       exit 1
     fi
   done
-  if [ -f "$REPO/userbot/k8s/active" ]; then
-    log "userbot panel hook"
-    if kubectl get secret userbot-common-secrets -n userbot >/dev/null 2>&1; then
-      echo "  userbot-common-secrets already present in userbot ns, not touching"
-    elif kubectl get secret userbot-common-secrets -n default >/dev/null 2>&1; then
-      echo "  bootstrapping userbot-common-secrets into userbot ns"
-      kubectl get secret userbot-common-secrets -n default -o json \
-        | jq 'del(.metadata.annotations,.metadata.creationTimestamp,.metadata.resourceVersion,.metadata.uid,.metadata.managedFields) | .metadata.namespace = "userbot"' \
-        | kubectl apply -f -
-    else
-      echo "  WARNING: userbot-common-secrets missing in both default and userbot ns; create it manually from the laptop"
-    fi
-  fi
   restart_stale_images
 
   # No verification here on purpose. This stage may be killed at any point by
@@ -825,6 +954,32 @@ smoke_hosts() {
     | sort -u
 }
 
+# Traefik's own list of the routes it actually built. The Kubernetes CRs are the
+# wrong source for this: when a middleware fails to load, Traefik drops the
+# router that referenced it and leaves the CR behind looking perfectly healthy.
+#
+# api.insecure is already on for the internal `traefik` entrypoint, but the pod
+# IP is not routable from the node, so read it through kubectl exec rather than
+# standing up a port-forward. HTTP only: the TCP routers match on HostSNI(`*`)
+# and the UDP ones carry no rule at all, both selected by entrypoint and port,
+# so neither can answer whether a given host has a route.
+traefik_http_routes() {
+  kubectl -n traefik exec deploy/traefik -- \
+    wget -qO- --timeout=10 http://127.0.0.1:8080/api/http/routers 2>/dev/null \
+    | jq -c '[.[] | {status, rule: (.rule // "")}]'
+}
+
+# The hosts Traefik currently routes to, one per line. Every backticked token of
+# an enabled rule counts, which is a superset of the hosts -- PathPrefix values
+# land here too, harmlessly -- but it keeps the host syntax in one place instead
+# of a matcher per host. The scan keeps the delimiters, so strip them: what
+# belongs in a comparison against a hostname is the bare name.
+traefik_routed_hosts() {
+  jq -r '[.[] | select(.status == "enabled") | (.rule // "")
+           | scan("`[^`]+`") | ltrimstr("`") | rtrimstr("`")]
+         | unique | .[]' <<<"$1"
+}
+
 # stage_verify_k8s watches the rollout, which reports that the pods converged.
 # It cannot tell a converged pod from a serving one: a route pointing at the
 # wrong port, a Service selector that matches nothing the app listens on, a 500
@@ -836,6 +991,12 @@ smoke_hosts() {
 # or a 404 from a path the service does not serve still means the chain is
 # intact. Only a transport failure (no DNS, refused, timeout) or a 5xx means
 # the service is not serving, and only those fail the run.
+#
+# Except that a 404 is not evidence on its own. A router Traefik refused to
+# build answers with the same 404 and nothing behind it, so a middleware that
+# fails to load takes down every route that referenced it while
+# this stage reports `ok` for all of them. No status code separates those two
+# cases, so ask Traefik which routes it built and fail on the difference.
 stage_smoke() {
   cd "$REPO"
   select_manifests >/dev/null
@@ -882,11 +1043,52 @@ stage_smoke() {
     esac
   done
 
+  # Second gate. The probe above only means something if a router matched the
+  # host in the first place, so compare the hosts we expect against the hosts
+  # Traefik reports and fail on the difference.
+  local routes routed
+  if ! routes="$(traefik_http_routes)"; then
+    echo "ERROR: could not read Traefik's router list, refusing to report success"
+    return 1
+  fi
+  routed="$(traefik_routed_hosts "$routes")"
+
+  local -a unrouted=()
+  local tries=3
+  while :; do
+    unrouted=()
+    for h in "${hosts[@]}"; do
+      grep -qxF "$h" <<<"$routed" || unrouted+=("$h")
+    done
+    if [ "${#unrouted[@]}" -eq 0 ]; then
+      break
+    fi
+    # A router mid-rollout is legitimately absent for a moment. A middleware
+    # that failed to load stays absent, so waiting cannot paper over it.
+    if [ "$tries" -le 1 ]; then
+      break
+    fi
+    tries=$((tries - 1))
+    warn "${#unrouted[@]} host(s) have no enabled route yet, re-checking in 10s"
+    sleep 10
+    if ! routes="$(traefik_http_routes)"; then
+      break
+    fi
+    routed="$(traefik_routed_hosts "$routes")"
+  done
+
+  if [ "${#unrouted[@]}" -ne 0 ]; then
+    for h in "${unrouted[@]}"; do
+      echo "  NO ROUTE $h (Traefik has no enabled router for this host)"
+    done
+    bad=1
+  fi
+
   if [ "$bad" -ne 0 ]; then
     echo "ERROR: at least one active service is not serving over its public route"
     return 1
   fi
-  echo "all ${#hosts[@]} route(s) answered"
+  echo "all ${#hosts[@]} route(s) answered and have a router"
 }
 
 stage_apply_compose() {

.gitea/workflows/deploy.yaml

@@ -39,10 +39,15 @@ env:
 
 jobs:
   preflight:
+    # Autodeploy defaults to OFF: pushes deploy only when the AUTODEPLOY repo
+    # variable is set to 'true' (Settings -> Actions -> Variables). A manual
+    # Run workflow always bypasses the switch: dispatching it is the explicit
+    # intent to deploy.
     if: >-
-      github.event_name != 'workflow_run' ||
+      (vars.AUTODEPLOY == 'true' || github.event_name == 'workflow_dispatch') &&
+      (github.event_name != 'workflow_run' ||
       (github.event.workflow_run.conclusion == 'success' &&
-       github.event.workflow_run.head_branch == 'main')
+       github.event.workflow_run.head_branch == 'main'))
     runs-on: [self-hosted, linux, arch, homelab, prod]
     timeout-minutes: 10
     steps:
@@ -73,8 +78,34 @@ jobs:
     needs: [validate]
     runs-on: [self-hosted, linux, arch, homelab, prod]
     # Apply only, no verification, so this is just the work itself: snapshot,
-    # then up to three sequential `helm upgrade --atomic --timeout 10m`, then the
-    # apply loop. Verification has its own job and its own budget.
+    # then sequential `helm upgrade --install --wait --rollback-on-failure --timeout 10m`, then the apply loop.
+    # Verification has its own job and its own budget.
+    #
+    # 45 is roughly four times the measured cost of the stage, which is
+    # deliberately not raised on a theory:
+    #
+    #   helm, healthy          3 no-op upgrades              ~3-5 min
+    #   helm, one release bad  rollback-on-failure spends its 10m,      ~10-15 min
+    #                            then rolls that one back
+    #   apply loop             ~40 manifests, 4 of which     ~1 min
+    #                           resolve an image digest
+    #   restart_stale_images   7.6s to find 8 workloads,     ~0.5 min
+    #                           9.8s to resolve their digests
+    #
+    # The helm figure is one release, not three: `set -e` aborts
+    # upgrade_helm_releases on the first failure, so a broken release costs
+    # 10m and the other two are never attempted. Multiplying 10m by three
+    # overstates the worst case by 20 minutes.
+    #
+    # The 45 minutes this was last raised to 45 were still not enough, and the
+    # job logs for those runs no longer exist, so what actually consumed the
+    # budget is not known - the two measurable candidates above account for
+    # ~15 of it. The unbounded `docker manifest inspect` against the registry's
+    # known hang mode is now bounded inside registry_digest (25s timeout, 3
+    # attempts): a dead registry fails each owned image after ~85s instead of
+    # hanging the stage, and a blinking one is retried instead of failing the
+    # whole apply file. Still open: make the stage announce which manifest it
+    # is working on, so a killed run leaves a diagnosable last line.
     timeout-minutes: 45
     steps:
       - name: Checkout repository
@@ -112,7 +143,22 @@ jobs:
       needs.apply-k8s.result != 'skipped' &&
       needs.apply-compose.result != 'skipped'
     runs-on: [self-hosted, linux, arch, homelab, prod]
-    # ceil(changed_workloads / 8) waves of ROLLOUT_TIMEOUT each, plus rollback.
+    # Not raised, because the arithmetic does not close.
+    #
+    # 32 workloads are under management and the wave width is 8, so the verify
+    # itself is 4 waves of ROLLOUT_TIMEOUT (300s) = 20 minutes worst case, when
+    # every rollout times out rather than converging. That is already 20 of 30.
+    #
+    # The other 10 would have to absorb rollback, and rollback_workloads is a
+    # serial `while read` loop at 300s per failed workload. 10 minutes buys two.
+    # Any larger number is buying a bigger multiple of an unbounded term rather
+    # than covering a known cost: 60 minutes buys eight, and 60 minutes is
+    # therefore not a bound, it is a guess with two digits.
+    #
+    # The number becomes derivable the moment rollback uses the same wave width
+    # as the verify: 32 failures then cost 4 waves = 20 minutes instead of 160,
+    # and 45 covers verify plus rollback at full width. That change is to the
+    # recovery path and is not folded into a timeout edit.
     timeout-minutes: 30
     steps:
       - name: Checkout repository

.gitea/workflows/install-ci-tools.sh

@@ -16,6 +16,10 @@ here="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 TOOLS_DIR="${TOOLS_DIR:-${RUNNER_TEMP:-/tmp}/homelab-tools}"
 BIN_DIR="$TOOLS_DIR/bin"
 mkdir -p "$BIN_DIR"
+# The just-installed tools must resolve inside this script too: callers only
+# prepend BIN_DIR to PATH after the script exits, so a bare `uv` below would
+# miss the binary install_uv just placed (exit 127 on a clean runner).
+export PATH="$BIN_DIR:$PATH"
 
 arch="$(uname -m)"
 # Upstream projects disagree on arch spelling: kubeconform and actionlint use
@@ -53,14 +57,31 @@ fetch() {
   fi
 }
 
+# resolve <command>
+# Absolute path to use for invoking a tool: the copy in BIN_DIR when present,
+# otherwise the name for PATH lookup. Every version check and every in-script
+# invocation goes through this, so a tool missing from both places reads as
+# "not installed" instead of dying with 127 under `set -e`.
+resolve() {
+  if [ -x "$BIN_DIR/$1" ]; then
+    printf '%s' "$BIN_DIR/$1"
+  else
+    printf '%s' "$1"
+  fi
+}
+
 # installed_version <command>
 # Prints the version of an already-installed tool, or nothing. Each tool spells
 # its version flag differently, hence the case.
 installed_version() {
-  local out
+  local bin out
+  bin="$(resolve "$1")"
+  if ! command -v "$bin" >/dev/null 2>&1; then
+    return 0
+  fi
   case "$1" in
-    kubeconform) out="$("$1" -v 2>/dev/null | head -1 || true)" ;;
-    *) out="$("$1" --version 2>/dev/null | head -1 || true)" ;;
+    kubeconform) out="$("$bin" -v 2>/dev/null | head -1 || true)" ;;
+    *) out="$("$bin" --version 2>/dev/null | head -1 || true)" ;;
   esac
   printf '%s' "$out"
 }
@@ -130,7 +151,7 @@ install_uv_tool() {
     return 0
   fi
   install_uv
-  UV_TOOL_BIN_DIR="$BIN_DIR" uv tool install --force "$1==$2" >/dev/null
+  UV_TOOL_BIN_DIR="$BIN_DIR" "$BIN_DIR/uv" tool install --force "$1==$2" >/dev/null
 }
 
 install_ruff() {

.gitea/workflows/renovate-run.yaml

@@ -81,7 +81,7 @@ jobs:
           docker run --rm \
             -v "$PWD/renovate/renovate.json:/opt/renovate/renovate.json:ro" \
             -e RENOVATE_PLATFORM=gitea \
-            -e RENOVATE_ENDPOINT=https://gitea.forust.xyz/api/v1 \
+            -e RENOVATE_ENDPOINT=https://git.forust.xyz/api/v1 \
             -e RENOVATE_TOKEN="$RENOVATE_TOKEN" \
             -e RENOVATE_GITHUB_COM_TOKEN="${RENOVATE_GITHUB_COM_TOKEN:-}" \
             -e RENOVATE_REPOSITORIES="${RENOVATE_REPOSITORIES:-forust/homelab}" \

.gitea/workflows/ssh-run.sh

@@ -21,10 +21,51 @@ ssh_key="$key_dir/deploy_key"
 printf '%s\n' "$DEPLOY_KEY" > "$ssh_key"
 chmod 600 "$ssh_key"
 
-ssh -i "$ssh_key" -p "$deploy_port" \
-  -o BatchMode=yes -o StrictHostKeyChecking=accept-new \
-  "${DEPLOY_USER}@${DEPLOY_HOST}" \
-  "REPO=$deploy_path APPLY_PRUNE=${APPLY_PRUNE:-false} DEPLOY_SHA=${DEPLOY_SHA:-} DEPLOY_SNAPSHOT_DIR=${DEPLOY_SNAPSHOT_DIR:-} STAGE=$1 bash -se" <<'EOF'
+# A connection that died silently used to hang until the job timeout, and the
+# stage was never re-run: one flaky TCP session cost a whole 45-minute apply.
+# ServerAlive* bounds how long a dead peer goes unnoticed, ConnectTimeout bounds
+# setup. Only exit 255 - ssh's own transport failures - is retried. A stage that
+# fails on its own merits exits with the remote's status, so a real failure
+# still surfaces its own log instead of burning three attempts. The stages are
+# declarative applies, so re-running one that had already committed is harmless.
+ssh_opts=(
+  -i "$ssh_key" -p "$deploy_port"
+  -o BatchMode=yes -o StrictHostKeyChecking=accept-new
+  -o ConnectTimeout=15
+  -o ServerAliveInterval=15 -o ServerAliveCountMax=4
+)
+
+rc=0
+# apply-k8s and apply-compose are separate workflow jobs so the graph stays
+# intact for the verify job, but on a single node they must not run at once:
+# host docker churn on top of cluster churn is what melts the node (load 40+,
+# netbird/ssh die, helm is left pending-*). Serialize them on the workstation
+# with a shared lock; whoever arrives second waits.
+remote_cmd=(bash -se)
+case "$1" in
+  apply-k8s | apply-compose)
+    remote_cmd=(flock -w 5400 /tmp/homelab-apply.lock bash -se)
+    ;;
+esac
+for attempt in 1 2 3; do
+  if [ "$attempt" -gt 1 ]; then
+    echo ":: warning::ssh transport failed, retrying (${attempt}/3)"
+    sleep $((attempt * 5))
+  fi
+  rc=0
+  # shellcheck disable=SC2029 # remote_cmd/ssh_opts expand on the client on purpose: they select the local ssh invocation, only the heredoc runs remotely.
+  ssh "${ssh_opts[@]}" "${DEPLOY_USER}@${DEPLOY_HOST}" \
+    env "REPO=$deploy_path" "APPLY_PRUNE=${APPLY_PRUNE:-false}" \
+    "DEPLOY_SHA=${DEPLOY_SHA:-}" "DEPLOY_SNAPSHOT_DIR=${DEPLOY_SNAPSHOT_DIR:-}" \
+    "STAGE=$1" "${remote_cmd[@]}" <<'EOF' || rc=$?
 source "$REPO/.gitea/workflows/deploy-lib.sh"
 run_stage "$STAGE"
 EOF
+  [ "$rc" -eq 0 ] && break
+  [ "$rc" -ne 255 ] && break
+done
+
+if [ "$rc" -ne 0 ]; then
+  echo ":: error::stage $1 failed over ssh (exit $rc)"
+fi
+exit "$rc"

adguardhome/k8s/adguard.yaml

@@ -58,6 +58,8 @@ spec:
   selector:
     matchLabels:
       app: adguard
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -73,7 +75,7 @@ spec:
               memory: "1.5Gi"
               cpu: "300m"
             requests:
-              memory: "500Mi"
+              memory: "512Mi"
               cpu: "50m"
           ports:
             - containerPort: 3000
@@ -82,6 +84,13 @@ spec:
               name: dns
             - containerPort: 853
               name: dot
+          readinessProbe:
+            tcpSocket:
+              port: dns
+            initialDelaySeconds: 5
+            periodSeconds: 5
+            successThreshold: 1
+            failureThreshold: 3
           volumeMounts:
             - name: adguard-data
               mountPath: /opt/adguardhome/work

adguardhome/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`dns.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: adguard-service
           port: 3000

authentik/k8s/authentik.yaml

@@ -34,6 +34,8 @@ spec:
   selector:
     matchLabels:
       app: authentik-server
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -52,8 +54,8 @@ spec:
             - containerPort: 9000
           resources:
             requests:
-              memory: "700Mi"
-              cpu: "300m"
+              memory: "768Mi"
+              cpu: "100m"
             limits:
               memory: "1.5Gi"
               cpu: "1000m"
@@ -68,6 +70,8 @@ spec:
   selector:
     matchLabels:
       app: authentik-worker
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -86,8 +90,8 @@ spec:
                 name: authentik-secrets
           resources:
             requests:
-              memory: "512Mi"
-              cpu: "300m"
+              memory: "320Mi"
+              cpu: "100m"
             limits:
-              memory: "1Gi"
+              memory: "768Mi"
               cpu: "700m"

authentik/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`auth.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: authentik-server-service
           port: 9000

cfddns/k8s/deployment.yaml

@@ -9,6 +9,8 @@ spec:
   selector:
     matchLabels:
       app: cfddns
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -22,10 +24,10 @@ spec:
           imagePullPolicy: Always
           resources:
             requests:
-              memory: "20Mi"
+              memory: "32Mi"
               cpu: "30m"
             limits:
-              memory: "64Mi"
+              memory: "128Mi"
               cpu: "50m"
           envFrom:
             - secretRef:

checkmk/k8s/checkmk.yaml

@@ -24,6 +24,8 @@ spec:
   selector:
     matchLabels:
       app: checkmk
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:

checkmk/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`cmk.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: checkmk-service
           port: 5000

cloudflared/k8s/deployment.yaml

@@ -9,6 +9,8 @@ spec:
   selector:
     matchLabels:
       app: cloudflared
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -30,8 +32,8 @@ spec:
                   key: TUNNEL_TOKEN
           resources:
             requests:
-              memory: "32Mi"
+              memory: "128Mi"
               cpu: "30m"
             limits:
-              memory: "128Mi"
+              memory: "256Mi"
               cpu: "200m"

converters/compose.yaml

@@ -1,7 +1,7 @@
 services:
   convertx:
     container_name: convertx
-    image: ghcr.io/c4illin/convertx:v0.18.0
+    image: ghcr.io/c4illin/convertx:v0.19.0
     restart: unless-stopped
     ports:
       - "9992:3000"

converters/k8s/bentopdf.yaml

@@ -31,12 +31,13 @@ spec:
           name: bentopdf
           ports:
             - containerPort: 8080
+          # p95 4M, max 11M over 7 days. Was 50Mi/700Mi.
           resources:
             requests:
-              memory: "50Mi"
+              memory: "32Mi"
               cpu: "50m"
               ephemeral-storage: "100Mi"
             limits:
-              memory: "700Mi"
+              memory: "128Mi"
               cpu: "700m"
               ephemeral-storage: "5Gi"

converters/k8s/convertx.yaml

@@ -20,13 +20,15 @@ spec:
   selector:
     matchLabels:
       app: convertx
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
         app: convertx
     spec:
       containers:
-        - image: ghcr.io/c4illin/convertx:v0.18.0
+        - image: ghcr.io/c4illin/convertx:v0.19.0
           name: convertx
           envFrom:
             - configMapRef:
@@ -38,13 +40,14 @@ spec:
           volumeMounts:
             - mountPath: /data
               name: data
+          # p95 85M, max 136M over 7 days, spikes while converting. Was 250Mi/1.5Gi.
           resources:
             requests:
-              memory: "250Mi"
+              memory: "128Mi"
               cpu: "100m"
             limits:
               cpu: "1500m"
-              memory: "1.5Gi"
+              memory: "512Mi"
       volumes:
         - name: data
           persistentVolumeClaim:

crowdsec/k8s/crowdsec-middleware.yaml

@@ -1,24 +0,0 @@
-apiVersion: traefik.io/v1alpha1
-kind: Middleware
-metadata:
-  name: crowdsec-bouncer
-  namespace: crowdsec
-spec:
-  plugin:
-    crowdsec-bouncer:
-      enabled: true
-      LogLevel: INFO
-      CrowdsecMode: live
-      CrowdsecLapiScheme: http
-      CrowdsecLapiHost: crowdsec-service.crowdsec.svc.cluster.local:8080
-      CrowdsecLapiKeyFile: "/etc/traefik/secrets/traefik-api-key"
-      # LAPI lookup is SYNCHRONOUS and per-request: the plugin blocks on
-      # `GET /v1/decisions?ip=...&banned=true` before the request reaches
-      # the backend, and fails CLOSED (403) if the lookup exceeds the
-      # timeout. Unset, the fork defaults to 10s, which is an eternity for
-      # a request path: a single slow LAPI (idle 1.3-7.4s here) turned
-      # every request into a 10s hang and then a self-inflicted 403.
-      # 2s keeps the fail-closed path fast and bounded; with the LAPI
-      # resourced properly (see crowdsec-values.yaml) the lookup is
-      # sub-100ms and this budget is never hit.
-      CrowdsecLapiTimeout: "2s"

crowdsec/k8s/crowdsec-values.yaml

@@ -16,6 +16,12 @@ agent:
       value: crowdsecurity/traefik crowdsecurity/base-http-scenarios
     - name: DISABLE_COLLECTIONS
       value: crowdsecurity/sshd
+    # Bans on 401/403 bursts hurt more than they protect: with L3 enforcement
+    # a false positive cuts the IP off everything (SSH included), and past
+    # incidents show legit automation (deploy runner, mesh peers, registry
+    # pulls) tripping this probe. Probing/XSS/SQLi/CVE scenarios stay.
+    - name: DISABLE_SCENARIOS
+      value: crowdsecurity/http-generic-bf
   metrics:
     enabled: true
     serviceMonitor:
@@ -58,26 +64,15 @@ config:
           reason: "Mobile IP whitelist"
           cidr:
             - "84.245.64.0/18"
-
-  postoverflows:
-    s01-whitelist:
-      home-dynamic-ip.yaml: |
-        name: forust/home-dynamic-ip
-        description: "Whitelist home dynamic IP"
-        whitelist:
-          reason: "Home dynamic IP"
-          expression:
-            - evt.Overflow.Alert.Source.IP in LookupHost("ddns.forust.xyz")
-      # The hairpin-NAT address of the router (192.168.88.1) is what the
-      # Gitea Actions runner presents to Traefik - it is NOT the home
-      # dynamic IP, so the whitelist above did not cover it. During a
-      # deploy the runner POSTs to the Actions API many times a second;
-      # a single 403 storm was enough to earn it a 4h ban and break every
-      # later job. Whitelisting the whole LAN also covers phones and
-      # tablets browsing over 192.168.88.0/24.
-      lan.yaml: |
-        name: forust/lan
-        description: "Whitelist local network"
+      # CrowdSec's own guidance: CIDR allowlisting belongs at the parser stage.
+      # A parser whitelist discards the event before it reaches a bucket, so
+      # these addresses never produce an overflow and never become a decision.
+      # A postoverflow whitelist is checked only *after* the ban exists, and
+      # the bouncer answers 403 for as long as it does - which is a window we
+      # do not want the deploy sitting in.
+      local-network.yaml: |
+        name: forust/local-network
+        description: "Whitelist loopback, private and VPN networks"
         whitelist:
           reason: "Local network"
           cidr:
@@ -85,6 +80,88 @@ config:
             - "10.0.0.0/8"
             - "172.16.0.0/12"
             - "192.168.0.0/16"
+            # CGNAT range (RFC 6598). The workstation and the k0s node live
+            # here on WireGuard, and 100.64.0.0/10 is not covered by the
+            # RFC 1918 blocks above.
+            - "100.64.0.0/10"
+            - "169.254.0.0/16"
+            - "fc00::/7"
+            - "fe80::/10"
+      vps-whitelist.yaml: |
+        name: forust/vps-whitelist
+        description: "Whitelist static VPS"
+        whitelist:
+          reason: "VPS"
+          ip:
+            - "193.181.211.79"
+
+  postoverflows:
+    s01-whitelist:
+      # The one whitelist that has to stay here: resolving a hostname is a
+      # network call, and the docs put expensive lookups in postoverflows on
+      # purpose - it runs only when a bucket actually overflows.
+      # ddns.forust.xyz is the public home address, not a private one, so
+      # forust/local-network does not cover it.
+      home-dynamic-ip.yaml: |
+        name: forust/home-dynamic-ip
+        description: "Whitelist home dynamic IP"
+        whitelist:
+          reason: "Home dynamic IP"
+          expression:
+            - evt.Overflow.Alert.Source.IP in LookupHost("ddns.forust.xyz")
+
+  # LAPI-only main config override, merged over config.yaml. NOTE: the
+  # chart's own default for this key is REPLACED, not merged, so its
+  # auto_registration block is repeated verbatim below - drop it and the
+  # agent can no longer register itself.
+  config.yaml.local: |
+    api:
+      server:
+        auto_registration: # Activate if not using TLS for authentication
+          enabled: true
+          token: "${REGISTRATION_TOKEN}" # /!\ Do not modify this variable (auto-generated and handled by the chart)
+          allowed_ranges: # /!\ Make sure to adapt to the pod IP ranges used by your cluster
+            - "127.0.0.1/32"
+            - "192.168.0.0/16"
+            - "10.0.0.0/8"
+            - "172.16.0.0/12"
+        # This homelab has no egress to console.crowdsec.cloud: DNS does
+        # not resolve. The LAPI kept trying anyway ("Signal push: N
+        # signals to push", "capi metrics: sending" every 10s) and each
+        # attempt sat on a resolver timeout WHILE HOLDING A WRITE
+        # TRANSACTION, which is what kept stalling per-request decision
+        # lookups even with WAL enabled. Nothing to share and nothing to
+        # pull - turn the Central API off instead of letting it block the
+        # only database writer we have.
+        online_client:
+          sharing: false
+          pull:
+            community: false
+            blocklists: false
+        disable_usage_metrics_export: true
+    db_config:
+      # SQLite without WAL serialises every reader behind the writer's
+      # rollback journal, and the LAPI writes constantly: the agent pushes
+      # Traefik alerts read from Loki, the metrics collector counts
+      # decisions, the bouncer touches "last pull" on every request.
+      # Symptom: decision lookups taking 10-30s (and a second connection
+      # that could not even open the database) while the LAPI sat at 28m
+      # CPU - the process was blocked in fsync, not computing. Every
+      # bouncer-protected request then blew through the plugin timeout and
+      # fail-closed with 403, on every site at once.
+      # The PVC is local-path-retain (hostPath), not a network share, so
+      # WAL is safe here; the crowdsec docs recommend it for exactly this
+      # ("allowing more concurrency in SQLite that will improve
+      # performances in most scenarios").
+      use_wal: true
+      # Keeps the alert table bounded. At the 5000/7d default the file
+      # reached 54MB in 15 days off the Traefik access log alone, and the
+      # metrics collector counts decisions on a timer; a smaller working
+      # set means fewer full scans. Crowdsec only prunes - SQLite never
+      # shrinks the file, so the size stays until a manual VACUUM.
+      flush:
+        max_items: 1000
+        max_age: 24h
 
 lapi:
   env:

crowdsec/k8s/janitor-cronjob.yaml

@@ -30,10 +30,14 @@
 #       3. ensure the static machine exists, recreating it with the
 #          Secret password if missing (agent retry loops reconnect
 #          on their own - same name + same password);
-#       4. prune bouncer entries idle for 30d;
-#       5. delete decisions from LePresidente/http-generic-403-bf, a hub
-#          scenario that bans an IP for 4h after 5 POST-403s in 10s and
-#          therefore bans us for our own bouncer's fail-closed 403s.
+#       4. prune bouncer entries idle for 30d.
+#
+# It used to also delete LePresidente/http-generic-403-bf decisions hourly.
+# That was a workaround for the bouncer failing closed on a slow LAPI and
+# 403-ing the deploy runner into a 4h ban. The bouncer now polls decisions
+# into a cache and never blocks on an unreachable LAPI, so it cannot
+# manufacture those 403s any more, and the scenario only fires against real
+# scanners - deleting their decisions hourly was undoing a working ban.
 #
 # Manual apply (crowdsec/k8s is NOT managed by deploy.yaml):
 #   kubectl apply -f crowdsec/k8s/janitor-cronjob.yaml
@@ -196,19 +200,3 @@ spec:
                   fi
                   echo "== 4. prune stale bouncers (no pull for 30d) =="
                   $LAPI_EXEC cscli bouncers prune -d 720h --force
-                  echo "== 5. drop http-403-bf decisions (4h self-bans) =="
-                  # `LePresidente/http-generic-403-bf` (hub item
-                  # crowdsecurity/http-generic-bf v0.9) bans any source IP
-                  # after 5 POSTs answered 403 within 10s, for 4h. That
-                  # includes 403s this homelab generates ITSELF (any
-                  # bouncer fail-closed, any app CSRF/rate-limit 403), and a
-                  # 4h ban on the runner/home IP silently breaks deploys and
-                  # browsing. The scenario cannot be removed per-scenario -
-                  # it is baked into a hub item, and disabling the whole
-                  # base-http-scenarios collection would drop ~40 useful
-                  # detections. Instead we keep the detection and drop its
-                  # decisions hourly; the LAN/home whitelists in
-                  # crowdsec-values.yaml handle the legit sources, so this
-                  # only ever hits real scanners (who are re-banned anyway).
-                  $LAPI_EXEC cscli decisions delete \
-                    --scenario LePresidente/http-generic-403-bf --all || true

dockmon/k8s/ingress.yaml

@@ -18,8 +18,6 @@ spec:
     - match: Host(`dockmon.forust.xyz`)
       kind: Rule
       middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
         - name: security-headers@file
       services:
         - name: dockmon-service

downtify/compose.yaml

@@ -1,7 +1,7 @@
 services:
   downtify:
     container_name: downtify
-    image: ghcr.io/henriquesebastiao/downtify:3.1.0
+    image: ghcr.io/henriquesebastiao/downtify:3.4.0
     restart: unless-stopped
     # ports:
     #   - '7077:8000'

downtify/k8s/downtify.yaml

@@ -20,6 +20,8 @@ spec:
   selector:
     matchLabels:
       app: downtify
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -27,7 +29,7 @@ spec:
     spec:
       containers:
         - name: downtify
-          image: ghcr.io/henriquesebastiao/downtify:3.1.0
+          image: ghcr.io/henriquesebastiao/downtify:3.4.0
           ports:
             - containerPort: 8000
           volumeMounts:

downtify/k8s/ingress.yaml

@@ -10,8 +10,6 @@ spec:
     - match: Host(`downtify.forust.xyz`)
       kind: Rule
       middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
         - name: security-chain@file
       services:
         - name: downtify-service

dtek_notif/Dockerfile

@@ -1,13 +0,0 @@
-FROM python:3.9-alpine
-
-WORKDIR /app
-
-# Установка зависимостей
-COPY requirements.txt .
-RUN pip install --no-cache-dir -r requirements.txt
-
-# Копирование кода
-COPY main.py .
-COPY .env .
-# Запуск бота
-CMD ["python", "-u", "main.py"]

dtek_notif/LICENSE

@@ -1,373 +0,0 @@
-Mozilla Public License Version 2.0
-==================================
-
-1. Definitions
---------------
-
-1.1. "Contributor"
-    means each individual or legal entity that creates, contributes to
-    the creation of, or owns Covered Software.
-
-1.2. "Contributor Version"
-    means the combination of the Contributions of others (if any) used
-    by a Contributor and that particular Contributor's Contribution.
-
-1.3. "Contribution"
-    means Covered Software of a particular Contributor.
-
-1.4. "Covered Software"
-    means Source Code Form to which the initial Contributor has attached
-    the notice in Exhibit A, the Executable Form of such Source Code
-    Form, and Modifications of such Source Code Form, in each case
-    including portions thereof.
-
-1.5. "Incompatible With Secondary Licenses"
-    means
-
-    (a) that the initial Contributor has attached the notice described
-        in Exhibit B to the Covered Software; or
-
-    (b) that the Covered Software was made available under the terms of
-        version 1.1 or earlier of the License, but not also under the
-        terms of a Secondary License.
-
-1.6. "Executable Form"
-    means any form of the work other than Source Code Form.
-
-1.7. "Larger Work"
-    means a work that combines Covered Software with other material, in 
-    a separate file or files, that is not Covered Software.
-
-1.8. "License"
-    means this document.
-
-1.9. "Licensable"
-    means having the right to grant, to the maximum extent possible,
-    whether at the time of the initial grant or subsequently, any and
-    all of the rights conveyed by this License.
-
-1.10. "Modifications"
-    means any of the following:
-
-    (a) any file in Source Code Form that results from an addition to,
-        deletion from, or modification of the contents of Covered
-        Software; or
-
-    (b) any new file in Source Code Form that contains any Covered
-        Software.
-
-1.11. "Patent Claims" of a Contributor
-    means any patent claim(s), including without limitation, method,
-    process, and apparatus claims, in any patent Licensable by such
-    Contributor that would be infringed, but for the grant of the
-    License, by the making, using, selling, offering for sale, having
-    made, import, or transfer of either its Contributions or its
-    Contributor Version.
-
-1.12. "Secondary License"
-    means either the GNU General Public License, Version 2.0, the GNU
-    Lesser General Public License, Version 2.1, the GNU Affero General
-    Public License, Version 3.0, or any later versions of those
-    licenses.
-
-1.13. "Source Code Form"
-    means the form of the work preferred for making modifications.
-
-1.14. "You" (or "Your")
-    means an individual or a legal entity exercising rights under this
-    License. For legal entities, "You" includes any entity that
-    controls, is controlled by, or is under common control with You. For
-    purposes of this definition, "control" means (a) the power, direct
-    or indirect, to cause the direction or management of such entity,
-    whether by contract or otherwise, or (b) ownership of more than
-    fifty percent (50%) of the outstanding shares or beneficial
-    ownership of such entity.
-
-2. License Grants and Conditions
---------------------------------
-
-2.1. Grants
-
-Each Contributor hereby grants You a world-wide, royalty-free,
-non-exclusive license:
-
-(a) under intellectual property rights (other than patent or trademark)
-    Licensable by such Contributor to use, reproduce, make available,
-    modify, display, perform, distribute, and otherwise exploit its
-    Contributions, either on an unmodified basis, with Modifications, or
-    as part of a Larger Work; and
-
-(b) under Patent Claims of such Contributor to make, use, sell, offer
-    for sale, have made, import, and otherwise transfer either its
-    Contributions or its Contributor Version.
-
-2.2. Effective Date
-
-The licenses granted in Section 2.1 with respect to any Contribution
-become effective for each Contribution on the date the Contributor first
-distributes such Contribution.
-
-2.3. Limitations on Grant Scope
-
-The licenses granted in this Section 2 are the only rights granted under
-this License. No additional rights or licenses will be implied from the
-distribution or licensing of Covered Software under this License.
-Notwithstanding Section 2.1(b) above, no patent license is granted by a
-Contributor:
-
-(a) for any code that a Contributor has removed from Covered Software;
-    or
-
-(b) for infringements caused by: (i) Your and any other third party's
-    modifications of Covered Software, or (ii) the combination of its
-    Contributions with other software (except as part of its Contributor
-    Version); or
-
-(c) under Patent Claims infringed by Covered Software in the absence of
-    its Contributions.
-
-This License does not grant any rights in the trademarks, service marks,
-or logos of any Contributor (except as may be necessary to comply with
-the notice requirements in Section 3.4).
-
-2.4. Subsequent Licenses
-
-No Contributor makes additional grants as a result of Your choice to
-distribute the Covered Software under a subsequent version of this
-License (see Section 10.2) or under the terms of a Secondary License (if
-permitted under the terms of Section 3.3).
-
-2.5. Representation
-
-Each Contributor represents that the Contributor believes its
-Contributions are its original creation(s) or it has sufficient rights
-to grant the rights to its Contributions conveyed by this License.
-
-2.6. Fair Use
-
-This License is not intended to limit any rights You have under
-applicable copyright doctrines of fair use, fair dealing, or other
-equivalents.
-
-2.7. Conditions
-
-Sections 3.1, 3.2, 3.3, and 3.4 are conditions of the licenses granted
-in Section 2.1.
-
-3. Responsibilities
--------------------
-
-3.1. Distribution of Source Form
-
-All distribution of Covered Software in Source Code Form, including any
-Modifications that You create or to which You contribute, must be under
-the terms of this License. You must inform recipients that the Source
-Code Form of the Covered Software is governed by the terms of this
-License, and how they can obtain a copy of this License. You may not
-attempt to alter or restrict the recipients' rights in the Source Code
-Form.
-
-3.2. Distribution of Executable Form
-
-If You distribute Covered Software in Executable Form then:
-
-(a) such Covered Software must also be made available in Source Code
-    Form, as described in Section 3.1, and You must inform recipients of
-    the Executable Form how they can obtain a copy of such Source Code
-    Form by reasonable means in a timely manner, at a charge no more
-    than the cost of distribution to the recipient; and
-
-(b) You may distribute such Executable Form under the terms of this
-    License, or sublicense it under different terms, provided that the
-    license for the Executable Form does not attempt to limit or alter
-    the recipients' rights in the Source Code Form under this License.
-
-3.3. Distribution of a Larger Work
-
-You may create and distribute a Larger Work under terms of Your choice,
-provided that You also comply with the requirements of this License for
-the Covered Software. If the Larger Work is a combination of Covered
-Software with a work governed by one or more Secondary Licenses, and the
-Covered Software is not Incompatible With Secondary Licenses, this
-License permits You to additionally distribute such Covered Software
-under the terms of such Secondary License(s), so that the recipient of
-the Larger Work may, at their option, further distribute the Covered
-Software under the terms of either this License or such Secondary
-License(s).
-
-3.4. Notices
-
-You may not remove or alter the substance of any license notices
-(including copyright notices, patent notices, disclaimers of warranty,
-or limitations of liability) contained within the Source Code Form of
-the Covered Software, except that You may alter any license notices to
-the extent required to remedy known factual inaccuracies.
-
-3.5. Application of Additional Terms
-
-You may choose to offer, and to charge a fee for, warranty, support,
-indemnity or liability obligations to one or more recipients of Covered
-Software. However, You may do so only on Your own behalf, and not on
-behalf of any Contributor. You must make it absolutely clear that any
-such warranty, support, indemnity, or liability obligation is offered by
-You alone, and You hereby agree to indemnify every Contributor for any
-liability incurred by such Contributor as a result of warranty, support,
-indemnity or liability terms You offer. You may include additional
-disclaimers of warranty and limitations of liability specific to any
-jurisdiction.
-
-4. Inability to Comply Due to Statute or Regulation
----------------------------------------------------
-
-If it is impossible for You to comply with any of the terms of this
-License with respect to some or all of the Covered Software due to
-statute, judicial order, or regulation then You must: (a) comply with
-the terms of this License to the maximum extent possible; and (b)
-describe the limitations and the code they affect. Such description must
-be placed in a text file included with all distributions of the Covered
-Software under this License. Except to the extent prohibited by statute
-or regulation, such description must be sufficiently detailed for a
-recipient of ordinary skill to be able to understand it.
-
-5. Termination
---------------
-
-5.1. The rights granted under this License will terminate automatically
-if You fail to comply with any of its terms. However, if You become
-compliant, then the rights granted under this License from a particular
-Contributor are reinstated (a) provisionally, unless and until such
-Contributor explicitly and finally terminates Your grants, and (b) on an
-ongoing basis, if such Contributor fails to notify You of the
-non-compliance by some reasonable means prior to 60 days after You have
-come back into compliance. Moreover, Your grants from a particular
-Contributor are reinstated on an ongoing basis if such Contributor
-notifies You of the non-compliance by some reasonable means, this is the
-first time You have received notice of non-compliance with this License
-from such Contributor, and You become compliant prior to 30 days after
-Your receipt of the notice.
-
-5.2. If You initiate litigation against any entity by asserting a patent
-infringement claim (excluding declaratory judgment actions,
-counter-claims, and cross-claims) alleging that a Contributor Version
-directly or indirectly infringes any patent, then the rights granted to
-You by any and all Contributors for the Covered Software under Section
-2.1 of this License shall terminate.
-
-5.3. In the event of termination under Sections 5.1 or 5.2 above, all
-end user license agreements (excluding distributors and resellers) which
-have been validly granted by You or Your distributors under this License
-prior to termination shall survive termination.
-
-************************************************************************
-*                                                                      *
-*  6. Disclaimer of Warranty                                           *
-*  -------------------------                                           *
-*                                                                      *
-*  Covered Software is provided under this License on an "as is"       *
-*  basis, without warranty of any kind, either expressed, implied, or  *
-*  statutory, including, without limitation, warranties that the       *
-*  Covered Software is free of defects, merchantable, fit for a        *
-*  particular purpose or non-infringing. The entire risk as to the     *
-*  quality and performance of the Covered Software is with You.        *
-*  Should any Covered Software prove defective in any respect, You     *
-*  (not any Contributor) assume the cost of any necessary servicing,   *
-*  repair, or correction. This disclaimer of warranty constitutes an   *
-*  essential part of this License. No use of any Covered Software is   *
-*  authorized under this License except under this disclaimer.         *
-*                                                                      *
-************************************************************************
-
-************************************************************************
-*                                                                      *
-*  7. Limitation of Liability                                          *
-*  --------------------------                                          *
-*                                                                      *
-*  Under no circumstances and under no legal theory, whether tort      *
-*  (including negligence), contract, or otherwise, shall any           *
-*  Contributor, or anyone who distributes Covered Software as          *
-*  permitted above, be liable to You for any direct, indirect,         *
-*  special, incidental, or consequential damages of any character      *
-*  including, without limitation, damages for lost profits, loss of    *
-*  goodwill, work stoppage, computer failure or malfunction, or any    *
-*  and all other commercial damages or losses, even if such party      *
-*  shall have been informed of the possibility of such damages. This   *
-*  limitation of liability shall not apply to liability for death or   *
-*  personal injury resulting from such party's negligence to the       *
-*  extent applicable law prohibits such limitation. Some               *
-*  jurisdictions do not allow the exclusion or limitation of           *
-*  incidental or consequential damages, so this exclusion and          *
-*  limitation may not apply to You.                                    *
-*                                                                      *
-************************************************************************
-
-8. Litigation
--------------
-
-Any litigation relating to this License may be brought only in the
-courts of a jurisdiction where the defendant maintains its principal
-place of business and such litigation shall be governed by laws of that
-jurisdiction, without reference to its conflict-of-law provisions.
-Nothing in this Section shall prevent a party's ability to bring
-cross-claims or counter-claims.
-
-9. Miscellaneous
-----------------
-
-This License represents the complete agreement concerning the subject
-matter hereof. If any provision of this License is held to be
-unenforceable, such provision shall be reformed only to the extent
-necessary to make it enforceable. Any law or regulation which provides
-that the language of a contract shall be construed against the drafter
-shall not be used to construe this License against a Contributor.
-
-10. Versions of the License
----------------------------
-
-10.1. New Versions
-
-Mozilla Foundation is the license steward. Except as provided in Section
-10.3, no one other than the license steward has the right to modify or
-publish new versions of this License. Each version will be given a
-distinguishing version number.
-
-10.2. Effect of New Versions
-
-You may distribute the Covered Software under the terms of the version
-of the License under which You originally received the Covered Software,
-or under the terms of any subsequent version published by the license
-steward.
-
-10.3. Modified Versions
-
-If you create software not governed by this License, and you want to
-create a new license for such software, you may create and use a
-modified version of this License if you rename the license and remove
-any references to the name of the license steward (except to note that
-such modified license differs from this License).
-
-10.4. Distributing Source Code Form that is Incompatible With Secondary
-Licenses
-
-If You choose to distribute Source Code Form that is Incompatible With
-Secondary Licenses under the terms of this version of the License, the
-notice described in Exhibit B of this License must be attached.
-
-Exhibit A - Source Code Form License Notice
--------------------------------------------
-
-  This Source Code Form is subject to the terms of the Mozilla Public
-  License, v. 2.0. If a copy of the MPL was not distributed with this
-  file, You can obtain one at https://mozilla.org/MPL/2.0/.
-
-If it is not possible or desirable to put the notice in a particular
-file, then You may include the notice in a location (such as a LICENSE
-file in a relevant directory) where a recipient would be likely to look
-for such a notice.
-
-You may add additional accurate notices of copyright ownership.
-
-Exhibit B - "Incompatible With Secondary Licenses" Notice
----------------------------------------------------------
-
-  This Source Code Form is "Incompatible With Secondary Licenses", as
-  defined by the Mozilla Public License, v. 2.0.

dtek_notif/docker-compose.yaml

@@ -1,15 +0,0 @@
-services:
-  dtek_notif:
-    build:
-      context: .
-      dockerfile: Dockerfile
-    image: gcr.forust.xyz/forust/dtek-notif:latest
-    pull_policy: build
-    restart: unless-stopped
-    environment:
-      - TZ=Europe/Kyiv
-    dns:
-      - 1.1.1.1
-      - 8.8.8.8
-    networks:
-      - default

dtek_notif/main.py

@@ -1,748 +0,0 @@
-import asyncio
-import contextlib
-import logging
-import os
-from datetime import datetime, timedelta
-
-import requests
-from aiogram import Bot, Dispatcher
-from aiogram.filters import Command
-from aiogram.types import KeyboardButton, Message
-from aiogram.utils.keyboard import ReplyKeyboardBuilder
-from bs4 import BeautifulSoup
-from dotenv import load_dotenv
-
-# Загрузка переменных окружения
-load_dotenv()
-
-# Настройки
-TELEGRAM_TOKEN = os.getenv('TELEGRAM_TOKEN', 'YOUR_TOKEN_HERE')
-ALLOWED_CHAT_IDS = list(map(int, os.getenv('ALLOWED_CHAT_IDS', '').split(','))) if os.getenv('ALLOWED_CHAT_IDS') else []
-CHECK_INTERVAL = int(os.getenv('CHECK_INTERVAL', '120'))
-
-# Параметры для запроса
-VOE_CITY_ID = int(os.getenv('VOE_CITY_ID', 'VOE_CITY_ID'))
-VOE_STREET_ID = int(os.getenv('VOE_STREET_ID', 'VOE_STREET_ID'))
-VOE_HOUSE_ID = int(os.getenv('VOE_HOUSE_ID', 'VOE_HOUSE_ID'))
-
-# Настройка логирования
-logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(name)s - %(levelname)s - %(message)s')
-logger = logging.getLogger(__name__)
-
-# Глобальные переменные
-bot = Bot(token=TELEGRAM_TOKEN)
-dp = Dispatcher()
-last_schedule: list[dict] | None = None
-last_notification_time: dict[str, datetime] = {}
-
-
-# ============================================================================
-# УТИЛИТЫ
-# ============================================================================
-
-
-def format_time_duration(minutes: int) -> str:
-    """Форматирует время из минут в часы и минуты"""
-    hours = minutes // 60
-    mins = minutes % 60
-
-    if hours == 0:
-        return f'{mins}м'
-    elif mins == 0:
-        return f'{hours}ч'
-    return f'{hours}ч {mins}м'
-
-
-def get_day_statistics(day_blocks: list[dict]) -> dict[str, int]:
-    """Получает статистику по дню"""
-    total_minutes = 0
-    confirmed_minutes = 0
-    possible_minutes = 0
-
-    for block in day_blocks:
-        for half in [block['first_half'], block['second_half']]:
-            if half['status'] == 'off':
-                total_minutes += 30
-                if half['confirmed']:
-                    confirmed_minutes += 30
-                else:
-                    possible_minutes += 30
-
-    return {'total': total_minutes, 'confirmed': confirmed_minutes, 'possible': possible_minutes}
-
-
-# ============================================================================
-# ПАРСИНГ ДАННЫХ
-# ============================================================================
-
-
-def parse_html(html: str) -> list[dict]:
-    """Парсит HTML с графиком отключений (логика от 15.11.2024)"""
-    soup = BeautifulSoup(html, 'html.parser')
-    cells = soup.select('.disconnection-detailed-table-cell.cell')
-
-    schedule = []
-    current_hour = 0
-    current_day = 0
-
-    for cell in cells:
-        if 'legend' in cell.get('class', []) or 'head' in cell.get('class', []):
-            continue
-
-        cell_classes = cell.get('class', [])
-
-        # ПРоверка статуса отключения на весь час
-        full_hour_off = 'has_disconnection' in cell_classes and 'full_hour' in cell_classes
-
-        hour_block = cell.select_one('.hour_block')
-        if not hour_block:
-            continue
-
-        # Проверка подтверждённости отключения для всего часа
-        cell_confirmed = None
-        if 'confirm_1' in cell_classes:
-            cell_confirmed = True
-        elif 'confirm_0' in cell_classes:
-            cell_confirmed = False
-
-        # Проверка половин часа
-        left = hour_block.select_one('.half.left')
-        right = hour_block.select_one('.half.right')
-
-        def parse_half(half, is_full_hour_off: bool, cell_confirmed: bool | None = None) -> dict:
-            """Парсит половину часа"""
-            if not half:
-                return {'status': 'on', 'queue': None, 'confirmed': None}
-
-            half_classes = half.get('class', [])
-
-            # Если вся ячейка full_hour - используем статус ячейки
-            if is_full_hour_off:
-                return {'status': 'off', 'queue': None, 'confirmed': cell_confirmed}
-
-            # Определяем статус половины
-            if 'has_disconnection' in half_classes:
-                status = 'off'
-            elif 'no_disconnection' in half_classes:
-                status = 'on'
-            else:
-                status = 'on'  # По умолчанию считаем включенным
-
-            # Если выключено - ищем подробности
-            queue = None
-            confirmed = None
-
-            if status == 'off':
-                disconnection_div = half.select_one('.disconnection')
-                if disconnection_div:
-                    # Ищем номер черги в title
-                    if disconnection_div.has_attr('title'):
-                        title = disconnection_div['title']
-                        if 'Номер черги' in title or 'Номер черги:' in title:
-                            with contextlib.suppress(BaseException):
-                                queue = title.split(':')[-1].strip()
-
-                    # Определяем подтверждение
-                    disc_classes = disconnection_div.get('class', [])
-                    if 'disconnection_confirm_1' in disc_classes:
-                        confirmed = True
-                    elif 'disconnection_confirm_0' in disc_classes:
-                        confirmed = False
-
-            return {'status': status, 'queue': queue, 'confirmed': confirmed}
-
-        first_half_data = parse_half(left, full_hour_off, cell_confirmed)
-        second_half_data = parse_half(right, full_hour_off, cell_confirmed)
-
-        schedule.append(
-            {
-                'hour': current_hour,
-                'day': current_day,
-                'first_half': first_half_data,
-                'second_half': second_half_data,
-            }
-        )
-
-        current_hour += 1
-        if current_hour >= 24:
-            current_hour = 0
-            current_day += 1
-
-    return schedule
-
-
-def get_voe_html(city_id: int, street_id: int, house_id: int) -> str:
-    """Получает HTML с сайта VOE"""
-    url = 'https://www.voe.com.ua/disconnection/detailed?ajax_form=1&_wrapper_format=drupal_ajax'
-    headers = {
-        'Content-Type': 'application/x-www-form-urlencoded; charset=UTF-8',
-        'X-Requested-With': 'XMLHttpRequest',
-        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36',
-    }
-    data = {
-        'search_type': 0,
-        'city_id': city_id,
-        'street_id': street_id,
-        'house_id': house_id,
-        'form_build_id': 'form-Irv5aHw1R2FT_Ik2apyHOZ47hTH5xPNH_LQnBrmpSTc',
-        'form_id': 'disconnection_detailed_search_form',
-        '_triggering_element_name': 'search',
-        '_triggering_element_value': 'Показати',
-        '_drupal_ajax': 1,
-    }
-
-    try:
-        response = requests.post(url, headers=headers, data=data, timeout=10)
-        response.raise_for_status()
-        resp_json = response.json()
-
-        insert_html = next((item['data'] for item in resp_json if item.get('command') == 'insert'), None)
-
-        if not insert_html:
-            raise ValueError('HTML не найден в ответе')
-
-        return insert_html
-    except requests.exceptions.RequestException as e:
-        logger.error(f'Ошибка запроса VOE: {e}')
-        raise
-
-
-# ============================================================================
-# ФОРМАТИРОВАНИЕ СООБЩЕНИЙ
-# ============================================================================
-
-
-def get_main_keyboard():
-    """Создает главную клавиатуру"""
-    builder = ReplyKeyboardBuilder()
-    builder.row(KeyboardButton(text='📊 Графік'), KeyboardButton(text='🔄 Оновити'))
-    builder.row(KeyboardButton(text='📅 Сьогодні'), KeyboardButton(text='📅 Завтра'))
-    builder.row(KeyboardButton(text='ℹ️ Про бота'))
-    return builder.as_markup(resize_keyboard=True)
-
-
-def format_schedule_message(schedule: list[dict], days_to_show: int = 2) -> str:
-    """Форматирует полный график на несколько дней"""
-    lines = [
-        '⚡️ <b>Графік відключень світла</b>',
-        f'🕐 Оновлено: {datetime.now().strftime("%d.%m.%Y %H:%M:%S")}',
-        '─' * 30,
-        '',
-    ]
-
-    start_date = datetime.now()
-
-    for day in range(min(days_to_show, 2)):
-        day_blocks = [b for b in schedule if b['day'] == day]
-        if not day_blocks:
-            continue
-
-        date_str = (start_date + timedelta(days=day)).strftime('%d.%m.%Y')
-        day_name = '🌅 <b>Сьогодні</b>' if day == 0 else '🌄 <b>Завтра</b>'
-
-        lines.append(f'{day_name} ({date_str})')
-
-        # Статистика
-        stats = get_day_statistics(day_blocks)
-        if stats['total'] > 0:
-            lines.append(f'⏱ Всього: <code>{format_time_duration(stats["total"])}</code>')
-            if stats['confirmed'] > 0:
-                lines.append(f'🔴 Підтверджено: <code>{format_time_duration(stats["confirmed"])}</code>')
-            if stats['possible'] > 0:
-                lines.append(f'🟠 Можливо: <code>{format_time_duration(stats["possible"])}</code>')
-        else:
-            lines.append('🟢 <b>Відключень немає!</b>')
-
-        lines.append('')
-
-        # Детальный список отключений
-        disconnections = []
-        current_status = None
-        start_time = None
-        current_confirmed = None
-        current_queue = None
-
-        for block in day_blocks:
-            hour = block['hour']
-
-            for half_idx, half in enumerate([block['first_half'], block['second_half']]):
-                time_str = f'{hour:02d}:00' if half_idx == 0 else f'{hour:02d}:30'
-
-                if half['status'] == 'off':
-                    if current_status != 'off':
-                        start_time = time_str
-                        current_confirmed = half['confirmed']
-                        current_queue = half['queue']
-                    current_status = 'off'
-                else:
-                    if current_status == 'off':
-                        icon = '🔴' if current_confirmed else '🟠'
-                        queue_text = f' (Ч{current_queue})' if current_queue else ''
-                        disconnections.append(f'{icon} <code>{start_time} - {time_str}</code>{queue_text}')
-                    current_status = half['status']
-
-        # Если день закончился на отключении
-        if current_status == 'off':
-            icon = '🔴' if current_confirmed else '🟠'
-            queue_text = f' (Ч{current_queue})' if current_queue else ''
-            next_hour = (day_blocks[-1]['hour'] + 1) % 24
-            end_time = f'{next_hour:02d}:00'
-            disconnections.append(f'{icon} <code>{start_time} - {end_time}</code>{queue_text}')
-
-        if disconnections:
-            for idx, disc in enumerate(disconnections, 1):
-                lines.append(f'{idx}. {disc}')
-
-        lines.append('')
-
-    lines.append('<i>🔴 = підтверджено • 🟠 = можливо • 🟢 = світло</i>')
-
-    return '\n'.join(lines)
-
-
-def format_single_day_schedule(schedule: list[dict], day: int) -> str:
-    """Форматирует график на один день"""
-    day_blocks = [b for b in schedule if b['day'] == day]
-    if not day_blocks:
-        return '❌ Немає даних для цього дня'
-
-    start_date = datetime.now()
-    date_str = (start_date + timedelta(days=day)).strftime('%d.%m.%Y')
-    day_name = '🟠 <b>Сьогодні</b>' if day == 0 else '🔶 <b>Завтра</b>'
-
-    lines = [f'{day_name}  •  {date_str}', '']
-
-    # Статистика
-    lines.append('<b>📊 Статистика</b>')
-    stats = get_day_statistics(day_blocks)
-
-    if stats['total'] == 0:
-        lines.append('└ 🟢 <b>Відключень немає!</b>')
-    else:
-        total_time = format_time_duration(stats['total'])
-        lines.append(f'├ ⏱ Всього: <code>{total_time}</code>')
-
-        if stats['confirmed'] > 0:
-            confirmed_time = format_time_duration(stats['confirmed'])
-            lines.append(f'├ 🔴 Підтверджено: <code>{confirmed_time}</code>')
-
-        if stats['possible'] > 0:
-            possible_time = format_time_duration(stats['possible'])
-            lines.append(f'└ 🟠 Можливо: <code>{possible_time}</code>')
-        else:
-            lines.append('└ 🟢 Решта часу світло')
-
-    lines.append('')
-
-    # Детальный список отключений
-    disconnections = []
-    current_status = None
-    start_time = None
-    current_confirmed = None
-    current_queue = None
-
-    for block in day_blocks:
-        hour = block['hour']
-
-        for half_idx, half in enumerate([block['first_half'], block['second_half']]):
-            time_str = f'{hour:02d}:00' if half_idx == 0 else f'{hour:02d}:30'
-
-            if half['status'] == 'off':
-                if current_status != 'off':
-                    start_time = time_str
-                    current_confirmed = half['confirmed']
-                    current_queue = half['queue']
-                current_status = 'off'
-            else:
-                if current_status == 'off':
-                    icon = '🔴' if current_confirmed else '🟠'
-                    queue_text = f' (Ч.{current_queue})' if current_queue else ''
-                    disconnections.append(f'{icon} <code>{start_time} - {time_str}</code>{queue_text}')
-                current_status = half['status']
-
-    # Если день закончился на отключении
-    if current_status == 'off':
-        icon = '🔴' if current_confirmed else '🟠'
-        queue_text = f' (Ч.{current_queue})' if current_queue else ''
-        next_hour = (day_blocks[-1]['hour'] + 1) % 24
-        end_time = f'{next_hour:02d}:00'
-        disconnections.append(f'{icon} <code>{start_time} - {end_time}</code>{queue_text}')
-
-    if disconnections:
-        lines.append('<b>⚡️ Розклад відключень</b>')
-        for idx, disc in enumerate(disconnections, 1):
-            lines.append(f'{idx}. {disc}')
-
-    lines.append('')
-    lines.append('<i>🔴 підтверджено  •  🟠 можливо  •  🟢 світло</i>')
-
-    return '\n'.join(lines)
-
-
-def schedules_differ(old_schedule: list[dict] | None, new_schedule: list[dict] | None) -> bool:
-    """Проверяет отличия между графиками"""
-    if old_schedule is None or new_schedule is None:
-        return True
-
-    if len(old_schedule) != len(new_schedule):
-        return True
-
-    for old, new in zip(old_schedule, new_schedule, strict=False):
-        if old['day'] >= 2:
-            break
-
-        if old['first_half'] != new['first_half'] or old['second_half'] != new['second_half']:
-            return True
-
-    return False
-
-
-# ============================================================================
-# УВЕДОМЛЕНИЯ
-# ============================================================================
-
-
-async def send_to_all_users(message_text: str, parse_mode: str = 'HTML'):
-    """Отправляет сообщение всем пользователям"""
-    if not ALLOWED_CHAT_IDS:
-        logger.warning('Нет допущенных ID чатов для отправки уведомлений')
-        return
-
-    for chat_id in ALLOWED_CHAT_IDS:
-        try:
-            await bot.send_message(chat_id, message_text, parse_mode=parse_mode)
-            logger.info(f'✅ Сообщение отправлено пользователю {chat_id}')
-        except Exception as e:
-            logger.error(f'❌ Ошибка отправки пользователю {chat_id}: {e}')
-        await asyncio.sleep(0.5)
-
-
-async def check_schedule():
-    """Проверяет график и отправляет уведомления"""
-    global last_schedule
-
-    try:
-        logger.info('🔍 Проверка графика...')
-        html = get_voe_html(VOE_CITY_ID, VOE_STREET_ID, VOE_HOUSE_ID)
-        new_schedule = parse_html(html)
-
-        if schedules_differ(last_schedule, new_schedule):
-            logger.info('✨ Обнаружены изменения!')
-            message = format_schedule_message(new_schedule, days_to_show=2)
-
-            if last_schedule is not None:
-                await send_to_all_users(f'🔄 <b>Графік оновлено!</b>\n\n{message}')
-
-            last_schedule = new_schedule
-        else:
-            logger.info('✓ Графік без змін')
-
-    except Exception as e:
-        logger.error(f'❌ Ошибка при проверке графика: {e}')
-
-
-async def check_upcoming_disconnections():
-    """Проверяет предстоящие события и отправляет предупреждения за 5 минут"""
-    global last_notification_time
-
-    if last_schedule is None:
-        return
-
-    now = datetime.now()
-    today_blocks = [b for b in last_schedule if b['day'] == 0]
-
-    # Создаем список всех переходов (off -> on или on -> off)
-    transitions = []
-    prev_status = None
-
-    for block in today_blocks:
-        hour = block['hour']
-
-        for half_idx, half in enumerate([block['first_half'], block['second_half']]):
-            minute = 0 if half_idx == 0 else 30
-            time_str = f'{hour:02d}:{minute:02d}'
-
-            current_status = half['status']
-
-            # Если статус изменился - это переход
-            if prev_status is not None and prev_status != current_status:
-                transitions.append(
-                    {
-                        'hour': hour,
-                        'minute': minute,
-                        'time_str': time_str,
-                        'from_status': prev_status,
-                        'to_status': current_status,
-                        'confirmed': half.get('confirmed'),
-                        'queue': half.get('queue'),
-                    }
-                )
-
-            prev_status = current_status
-
-    # Проверяем переходы
-    for transition in transitions:
-        event_time = now.replace(hour=transition['hour'], minute=transition['minute'], second=0, microsecond=0)
-
-        time_until = (event_time - now).total_seconds() / 60
-        notification_key = f'{transition["hour"]}:{transition["minute"]}_{transition["to_status"]}'
-
-        # Если за 5 минут до события (±1 минута) и еще не отправляли
-        if 4 <= time_until <= 6:
-            # Проверяем, не отправляли ли уже уведомление сегодня
-            if notification_key in last_notification_time:
-                last_notif_time = last_notification_time[notification_key]
-                if last_notif_time.date() == now.date():
-                    continue  # Уже отправляли сегодня
-
-            # Переход на ОТКЛЮЧЕНИЕ (on -> off)
-            if transition['from_status'] == 'on' and transition['to_status'] == 'off':
-                icon = '🔴' if transition['confirmed'] else '🟠'
-                status = 'підтверджено' if transition['confirmed'] else 'можливе'
-                queue_info = f' (Черга {transition["queue"]})' if transition['queue'] else ''
-
-                warning = (
-                    f'⚠️ <b>УВАГА! ВІДКЛЮЧЕННЯ</b>\n\n'
-                    f'Через ~5 хвилин\n'
-                    f'Час: <code>{transition["time_str"]}</code>\n'
-                    f'Статус: {icon} {status}{queue_info}'
-                )
-
-                await send_to_all_users(warning)
-                last_notification_time[notification_key] = now
-                logger.info(f'📢 Відправлено попередження про ВІДКЛЮЧЕННЯ в {transition["time_str"]}')
-
-            # Переход на ВКЛЮЧЕНИЕ (off -> on)
-            elif transition['from_status'] == 'off' and transition['to_status'] == 'on':
-                warning = (
-                    f'✅ <b>УВАГА! ВКЛЮЧЕННЯ</b>\n\n'
-                    f'Через ~5 хвилин буде світло\n'
-                    f'Час: <code>{transition["time_str"]}</code>'
-                )
-
-                await send_to_all_users(warning)
-                last_notification_time[notification_key] = now
-                logger.info(f'📢 Відправлено попередження про ВКЛЮЧЕННЯ в {transition["time_str"]}')
-
-
-async def monitoring_loop():
-    """Основной цикл мониторинга"""
-    await check_schedule()
-
-    while True:
-        try:
-            await asyncio.sleep(CHECK_INTERVAL)
-            await check_schedule()
-            await check_upcoming_disconnections()
-        except Exception as e:
-            logger.error(f'Ошибка в цикле мониторинга: {e}')
-            await asyncio.sleep(5)
-
-
-# ============================================================================
-# ОБРАБОТЧИКИ КОМАНД
-# ============================================================================
-
-
-@dp.message(Command('start'))
-async def cmd_start(message: Message):
-    """Обработчик /start"""
-    if message.chat.id not in ALLOWED_CHAT_IDS:
-        await message.answer('❌ У вас немає доступу до цього бота.')
-        return
-
-    await message.answer(
-        '👋 <b>Ласкаво просимо!</b>\n\n'
-        '🤖 <b>Бот для моніторингу графіку відключень світла</b>\n\n'
-        '✨ <b>Можливості:</b>\n'
-        '• 📊 Перегляд графіку на сьогодні і завтра\n'
-        '• 🔔 Автоматичні сповіщення за 5 хвилин до подій\n'
-        '• 🔄 Моніторинг змін графіку\n\n'
-        'Використовуйте кнопки нижче 👇',
-        parse_mode='HTML',
-        reply_markup=get_main_keyboard(),
-    )
-
-
-@dp.message(lambda msg: msg.text == 'ℹ️ Про бота')
-async def cmd_info(message: Message):
-    """Показывает информацию о боте"""
-    if message.chat.id not in ALLOWED_CHAT_IDS:
-        return
-
-    await message.answer(
-        '<b>ℹ️ Про бота</b>\n\n'
-        '🚀 <b>Версія:</b> 2.2 (Стабільна)\n\n'
-        '📝 <b>Реліз-ноути:</b>\n'
-        '├ 15.11.2024: Адаптація під оновлену логіку сайту VOE\n'
-        '├ Виправлено парсинг half.left та half.right\n'
-        '├ Покращено визначення підтвердження відключень\n'
-        '└ Оптимізовано обробку статусу для всієї години\n\n'
-        '⚡ <b>Функціональність:</b>\n'
-        '├ Моніторинг графіку 24/7\n'
-        '├ Сповіщення за 5 хвилин\n'
-        '├ Детальна статистика дня\n'
-        '└ Красива візуалізація\n\n'
-        '🔐 <b>Безпека:</b> Використовуються .env файли\n'
-        '💾 <b>Джерело:</b> voe.com.ua',
-        parse_mode='HTML',
-        reply_markup=get_main_keyboard(),
-    )
-
-
-@dp.message(Command('schedule'))
-async def cmd_schedule(message: Message):
-    """Показывает полный график"""
-    if message.chat.id not in ALLOWED_CHAT_IDS:
-        await message.answer('❌ У вас немає доступу.')
-        return
-
-    try:
-        await message.answer('⏳ Завантаження графіку...')
-        html = get_voe_html(VOE_CITY_ID, VOE_STREET_ID, VOE_HOUSE_ID)
-        schedule = parse_html(html)
-        text = format_schedule_message(schedule, days_to_show=2)
-        await message.answer(text, parse_mode='HTML', reply_markup=get_main_keyboard())
-    except Exception as e:
-        await message.answer(f'❌ <b>Помилка:</b> {str(e)}', parse_mode='HTML', reply_markup=get_main_keyboard())
-
-
-@dp.message(Command('today'))
-async def cmd_today(message: Message):
-    """Показывает график на сегодня"""
-    if message.chat.id not in ALLOWED_CHAT_IDS:
-        await message.answer('❌ У вас немає доступу.')
-        return
-
-    try:
-        await message.answer('⏳ Завантаження графіку сьогодні...')
-        html = get_voe_html(VOE_CITY_ID, VOE_STREET_ID, VOE_HOUSE_ID)
-        schedule = parse_html(html)
-        text = format_single_day_schedule(schedule, 0)
-        await message.answer(text, parse_mode='HTML', reply_markup=get_main_keyboard())
-    except Exception as e:
-        await message.answer(f'❌ <b>Помилка:</b> {str(e)}', parse_mode='HTML', reply_markup=get_main_keyboard())
-
-
-@dp.message(Command('tomorrow'))
-async def cmd_tomorrow(message: Message):
-    """Показывает график на завтра"""
-    if message.chat.id not in ALLOWED_CHAT_IDS:
-        await message.answer('❌ У вас немає доступу.')
-        return
-
-    try:
-        await message.answer('⏳ Завантаження графіку завтра...')
-        html = get_voe_html(VOE_CITY_ID, VOE_STREET_ID, VOE_HOUSE_ID)
-        schedule = parse_html(html)
-        text = format_single_day_schedule(schedule, 1)
-        await message.answer(text, parse_mode='HTML', reply_markup=get_main_keyboard())
-        # await message.answer("❌ Функція тимчасово недоступна. Чекаємо на оновлення сайту", parse_mode="HTML", reply_markup=get_main_keyboard())
-    except Exception as e:
-        await message.answer(f'❌ <b>Помилка:</b> {str(e)}', parse_mode='HTML', reply_markup=get_main_keyboard())
-
-
-@dp.message(Command('check'))
-async def cmd_check(message: Message):
-    """Принудительная проверка графика"""
-    if message.chat.id not in ALLOWED_CHAT_IDS:
-        await message.answer('❌ У вас немає доступу.')
-        return
-
-    try:
-        await message.answer('🔄 <b>Перевіряю графік...</b>', parse_mode='HTML')
-        html = get_voe_html(VOE_CITY_ID, VOE_STREET_ID, VOE_HOUSE_ID)
-        new_schedule = parse_html(html)
-
-        prefix = (
-            '✅ <b>Знайдено зміни!</b>\n\n'
-            if schedules_differ(last_schedule, new_schedule)
-            else '✓ <b>Графік без змін</b>\n\n'
-        )
-        result = prefix + format_schedule_message(new_schedule, days_to_show=2)
-
-        await message.answer(result, parse_mode='HTML', reply_markup=get_main_keyboard())
-    except Exception as e:
-        await message.answer(f'❌ <b>Помилка:</b> {str(e)}', parse_mode='HTML', reply_markup=get_main_keyboard())
-
-
-@dp.message()
-async def handle_text(message: Message):
-    """Обработчик текстовых сообщений и кнопок"""
-    if message.chat.id not in ALLOWED_CHAT_IDS:
-        return
-
-    text = message.text
-
-    # Кнопка "Графік"
-    if text == '📊 Графік':
-        await cmd_schedule(message)
-
-    # Кнопка "Сьогодні"
-    elif text == '📅 Сьогодні':
-        await cmd_today(message)
-
-    # Кнопка "Завтра"
-    elif text == '📅 Завтра':
-        await cmd_tomorrow(message)
-
-    # Кнопка "Оновити"
-    elif text == '🔄 Оновити':
-        await cmd_check(message)
-
-    # Кнопка "Про бота"
-    elif text == 'ℹ️ Про бота':
-        await cmd_info(message)
-
-    # Неизвестная команда
-    else:
-        await message.answer(
-            '❓ <b>Команда не розпізнана</b>\n\n'
-            'Використовуйте кнопки на клавіатурі або команди:\n'
-            '/start • /today • /tomorrow • /schedule • /check',
-            parse_mode='HTML',
-            reply_markup=get_main_keyboard(),
-        )
-
-
-# ============================================================================
-# ГЛАВНАЯ ФУНКЦИЯ
-# ============================================================================
-
-
-async def main():
-    """Главная функция"""
-    logger.info('=' * 50)
-    logger.info('ЗАПУСК БОТА V2.2 (stable 2.2, 15.11.2025)')
-    logger.info('=' * 50)
-
-    if not TELEGRAM_TOKEN or os.getenv('TELEGRAM_TOKEN', 'YOUR_TOKEN_HERE') == TELEGRAM_TOKEN:
-        logger.error('❌ TELEGRAM_TOKEN не конфігурований! Напишіть токен в .env файл')
-        return
-
-    if not ALLOWED_CHAT_IDS:
-        logger.error('❌ ALLOWED_CHAT_IDS не конфігуровані! Напишіть ID в .env файл')
-        return
-
-    logger.info(f'📌 Allowed chat ids: {ALLOWED_CHAT_IDS}')
-    logger.info(f'⏱ Інтервал перевірки: {CHECK_INTERVAL} сек')
-    logger.info('=' * 50)
-
-    # Запускаем мониторинг
-    monitoring_task = asyncio.create_task(monitoring_loop())
-
-    try:
-        await dp.start_polling(bot)
-    except KeyboardInterrupt:
-        logger.info('⏹ Бот зупинений користувачем')
-    finally:
-        monitoring_task.cancel()
-        await bot.session.close()
-        logger.info('✓ Підключення закрито')
-
-
-if __name__ == '__main__':
-    try:
-        asyncio.run(main())
-    except KeyboardInterrupt:
-        logger.info('⏹ Завершено')

dtek_notif/pyproject.toml

@@ -1,7 +0,0 @@
-[project]
-name = "dtek-notif"
-version = "0.1.0"
-description = "Add your description here"
-readme = "README.md"
-requires-python = ">=3.13"
-dependencies = []

dtek_notif/requirements.txt

@@ -1,5 +0,0 @@
-requests>=2.31.0
-beautifulsoup4>=4.12.0
-aiogram>=3.3.0
-python-dotenv>=1.0.0
-aiohttp>=3.9.0

edu_master/.env.example

@@ -1,10 +1,11 @@
-EDU_LOGIN=your_edu_login_here
-EDU_PASSWORD=your_edu_password_here
-EDU_URL_LOGIN=https://edu.edu.vn.ua/user/login
-EDU_URL_VERIFY=https://edu.edu.vn.ua/course/userlist
-PHPSESSID_INTERVAL=10
+KEEPER_LOGIN=your_edu_login_here
+KEEPER_PASSWORD=your_edu_password_here
+EDU_URL_BASE=https://edu.edu.vn.ua
+EDU_URL_LOGIN=/user/login
+EDU_URL_COURSES=/course/userlist
+KEEPER_INTERVAL=10
 USER_AGENT="Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/142.0.0.0 Safari/537.36"
-WEBINAR_URL=https://edu.edu.vn.ua/webinar/useractive
+EDU_URL_WEBINAR=/webinar/useractive
 WEBINAR_CHECK_INTERVAL=60
 REDIS_HOST=redis
 REDIS_PORT=6379

edu_master/compose.yaml

@@ -17,7 +17,7 @@ services:
 
   session-keeper:
     build: ./phpsessid-bot
-    image: gcr.forust.xyz/forust/session-keeper:latest
+    image: gcr.forust.xyz/forust/session-keeper:prod
     pull_policy: build
     env_file: .env
     restart: unless-stopped
@@ -33,7 +33,7 @@ services:
 
   webinar-checker:
     build: ./webinar-checker
-    image: gcr.forust.xyz/forust/webinar-checker:latest
+    image: gcr.forust.xyz/forust/webinar-checker:prod
     pull_policy: build
     env_file: .env
     restart: unless-stopped

edu_master/k8s/playwright.yaml

@@ -10,6 +10,8 @@ spec:
   selector:
     matchLabels:
       app: edu-master-playwright
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -20,6 +22,15 @@ spec:
           # renovate: datasource=docker depName=mcr.microsoft.com/playwright versioning=docker
           image: mcr.microsoft.com/playwright:v1.56.0-jammy
           imagePullPolicy: IfNotPresent
+          # p95 412M, max 478M over 7 days, no limit before. Request is set at p95
+          # so the pod is not an eviction candidate; the limit stays above 2x the
+          # request because browser page lifetimes are unpredictable.
+          resources:
+            requests:
+              cpu: "200m"
+              memory: "416Mi"
+            limits:
+              memory: "1Gi"
           command:
             - npx
             - -y

edu_master/k8s/redis.yaml

@@ -28,10 +28,10 @@ spec:
           resources:
             requests:
               cpu: 25m
-              memory: 64Mi
+              memory: 32Mi
             limits:
               cpu: 250m
-              memory: 256Mi
+              memory: 128Mi
           readinessProbe:
             exec:
               command: ["redis-cli", "ping"]

edu_master/k8s/session-keeper.yaml

@@ -10,6 +10,8 @@ spec:
   selector:
     matchLabels:
       app: edu-master-session-keeper
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -38,10 +40,10 @@ spec:
           resources:
             requests:
               cpu: 25m
-              memory: 96Mi
+              memory: 32Mi
             limits:
               cpu: 250m
-              memory: 256Mi
+              memory: 128Mi
           readinessProbe:
             exec:
               command: ["/bin/sh", "-ec", "redis-cli -h redis EXISTS EDU_PHPSESSID | grep -q 1"]

edu_master/k8s/webinar-checker.yaml

@@ -10,6 +10,8 @@ spec:
   selector:
     matchLabels:
       app: edu-master-webinar-checker
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -67,7 +69,7 @@ spec:
           resources:
             requests:
               cpu: "50m"
-              memory: "128Mi"
+              memory: "192Mi"
             limits:
               cpu: "600m"
-              memory: "512Mi"
+              memory: "384Mi"

edu_master/phpsessid-bot/bot.py

@@ -48,17 +48,43 @@ def touch_success_file():
         logger.error(f'Failed to touch success file: {e}')
 
 
+def refresh_session(session, redis_client):
+    """Publish a verified cookie with a lifetime tied to the refresh interval."""
+    login_response = session.post(
+        URL_LOGIN, data={'login': LOGIN, 'password': PASSWORD}, allow_redirects=True, timeout=(10, 30)
+    )
+    login_response.raise_for_status()
+    verify_response = session.get(URL_VERIFY, allow_redirects=False, timeout=(10, 30))
+    if verify_response.status_code != 200:
+        logger.warning('Session verification failed (HTTP %s)', verify_response.status_code)
+        return False
+    phpsessid = session.cookies.get('PHPSESSID')
+    if not phpsessid:
+        logger.warning('Verified response did not provide a PHPSESSID cookie')
+        return False
+    redis_client.set('EDU_PHPSESSID', phpsessid, ex=INTERVAL * 120)
+    touch_success_file()
+    logger.info('Verified session saved to Redis')
+    return True
+
+
 def main():
+    if not LOGIN or not PASSWORD:
+        raise ValueError('KEEPER_LOGIN and KEEPER_PASSWORD must be set')
+    if INTERVAL <= 0:
+        raise ValueError('KEEPER_INTERVAL must be a positive number of minutes')
     logger.info('Starting Session Keeper Bot')
 
     # Connect to Redis
     try:
-        redis_client = redis.Redis(host=REDIS_HOST, port=REDIS_PORT, decode_responses=True)
+        redis_client = redis.Redis(
+            host=REDIS_HOST, port=REDIS_PORT, decode_responses=True, socket_connect_timeout=5, socket_timeout=5
+        )
         redis_client.ping()
         logger.info(f'Connected to Redis at {REDIS_HOST}:{REDIS_PORT}')
     except Exception as e:
         logger.error(f'Failed to connect to Redis: {e}')
-        return
+        raise
 
     session = requests.Session()
 
@@ -83,44 +109,7 @@ def main():
 
     while True:
         try:
-            logger.info('Attempting login...')
-
-            # Login payload
-            payload = {'login': LOGIN, 'password': PASSWORD}
-
-            # Perform Login
-            # Note: The user request shows a POST to /user/login with form data
-            # We need to make sure we handle the PHPSESSID correctly.
-            # If we already have a PHPSESSID, requests will send it.
-
-            login_response = session.post(URL_LOGIN, data=payload, allow_redirects=True)
-
-            logger.info(f'Login Response Status: {login_response.status_code}')
-            logger.info(f'Cookies after login: {session.cookies.get_dict()}')
-
-            # Verify Session
-            logger.info('Verifying session...')
-            verify_response = session.get(URL_VERIFY, allow_redirects=False)
-
-            logger.info(f'Verify Response Status: {verify_response.status_code}')
-
-            if verify_response.status_code == 200:
-                logger.info('Session verification SUCCESS (200 OK).')
-                touch_success_file()
-
-                # Save PHPSESSID to Redis
-                phpsessid = session.cookies.get('PHPSESSID')
-                if phpsessid:
-                    try:
-                        redis_client.set('EDU_PHPSESSID', phpsessid)
-                        logger.info(f'Saved PHPSESSID to Redis: {phpsessid}')
-                    except Exception as e:
-                        logger.error(f'Failed to save PHPSESSID to Redis: {e}')
-            elif verify_response.status_code == 302:
-                logger.warning('Session verification FAILED (302 Redirect). Session might be invalid.')
-            else:
-                logger.warning(f'Session verification returned unexpected status: {verify_response.status_code}')
-
+            refresh_session(session, redis_client)
         except Exception as e:
             logger.error(f'An error occurred: {e}')
 

errorpages/compose.yaml

@@ -1,7 +1,7 @@
 services:
   errorpage:
     build: .
-    image: gcr.forust.xyz/forust/error-pages:latest
+    image: gcr.forust.xyz/forust/error-pages:prod
     pull_policy: build
     container_name: error-pages
     restart: unless-stopped

errorpages/k8s/error-pages.yaml

@@ -20,6 +20,8 @@ spec:
   selector:
     matchLabels:
       app: error-pages
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -28,6 +30,13 @@ spec:
       containers:
         - name: error-pages
           image: gcr.forust.xyz/forust/error-pages:prod
+          # p95 6M, max 10M, no limit before.
+          resources:
+            requests:
+              cpu: "10m"
+              memory: "32Mi"
+            limits:
+              memory: "128Mi"
           ports:
             - containerPort: 80
           readinessProbe:

gitea/compose.yaml

@@ -1,6 +1,6 @@
 services:
   server:
-    image: docker.gitea.com/gitea:1.27.3
+    image: docker.gitea.com/gitea:28.0.0
     container_name: gitea
     restart: always
     environment:
@@ -13,9 +13,12 @@ services:
       - GITEA__database__PASSWD=gitea
       - GITEA__database__NAME=gitea
       # Server
-      - GITEA__server__ROOT_URL=https://gitea.forust.xyz
+      - GITEA__server__ROOT_URL=https://git.forust.xyz
       - GITEA__server__SSH_DOMAIN=gitssh.forust.xyz
       - GITEA__server__SSH_PORT=2221
+      # Pin 28.0 defaults explicitly (see k8s/config.yaml for rationale)
+      - GITEA__service__DISABLE_REGISTRATION=true
+      - GITEA__actions__RUN_RETENTION_DAYS=90
       # Mailer
       - GITEA__mailer__ENABLED=true
       - GITEA__mailer__FROM=${SERVICE_EMAIL}
@@ -34,7 +37,7 @@ services:
       - "traefik.http.services.gitea.loadbalancer.server.port=3000"
 
       # Prod Router
-      - "traefik.http.routers.gitea.rule=Host(`gitea.forust.xyz`)"
+      - "traefik.http.routers.gitea.rule=Host(`git.forust.xyz`) || Host(`gitea.forust.xyz`)"
       - "traefik.http.routers.gitea.entrypoints=websecure"
       - "traefik.http.routers.gitea.tls.certresolver"
       # Local Router

gitea/k8s/certificates.yaml

@@ -8,6 +8,7 @@ spec:
   dnsNames:
     - gcr.forust.xyz
     - gitea.forust.xyz
+    - git.forust.xyz
   issuerRef:
     name: letsencrypt-prod
     kind: ClusterIssuer

gitea/k8s/config.yaml

@@ -4,11 +4,14 @@ metadata:
   name: gitea-config
   namespace: gitea
 data:
-  GITEA__server__DOMAIN: "gitea.forust.xyz"
-  GITEA__server__ROOT_URL: "https://gitea.forust.xyz"
+  GITEA__server__ROOT_URL: "https://git.forust.xyz"
   GITEA__server__SSH_DOMAIN: "gitssh.forust.xyz"
   GITEA__server__SSH_PORT: "2221"
 
+  GITEA__service__DISABLE_REGISTRATION: "true"
+
+  GITEA__actions__RUN_RETENTION_DAYS: "90"
+
   GITEA__database__DB_TYPE: "postgres"
   GITEA__database__HOST: "postgres.database.svc.cluster.local:5432"
   GITEA__database__NAME: "gitea"
@@ -17,6 +20,12 @@ data:
 
   GITEA__mailer__ENABLED: "false"
 
+  # No code/issue search needed: bleve reindexes the whole issue index on
+  # every pod restart (cron.rebuild_issue_indexer RUN_AT_START) and hammers
+  # the rotational disk for an hour. "db" serves issue search from postgres.
+  GITEA__indexer__ISSUE_INDEXER_TYPE: "db"
+  GITEA__indexer__REPO_INDEXER_ENABLED: "false"
+
   GITEA__log__logger.access.MODE: "console, file"
   USER_UID: "1000"
   USER_GID: "1000"

gitea/k8s/gitea.yaml

@@ -24,6 +24,8 @@ spec:
   selector:
     matchLabels:
       app: gitea
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -31,7 +33,7 @@ spec:
     spec:
       containers:
         - name: gitea
-          image: gitea/gitea:1.27.3
+          image: gitea/gitea:28.0.0
           envFrom:
             - configMapRef:
                 name: gitea-config
@@ -47,10 +49,10 @@ spec:
               mountPath: /data
           resources:
             requests:
-              memory: "512Mi"
-              cpu: "300m"
+              memory: "320Mi"
+              cpu: "100m"
             limits:
-              memory: "1.5Gi"
+              memory: "1Gi"
               cpu: "1300m"
       volumes:
         - name: gitea-data

gitea/k8s/ingress.yaml

@@ -7,24 +7,11 @@ spec:
   entryPoints:
     - websecure
   routes:
-    - match: Host(`gitea.forust.xyz`)
+    - match: Host(`gitea.forust.xyz`) || Host(`git.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: gitea-service
           port: 3000
-    # Registry route: NO crowdsec-bouncer.
-    # The bouncer plugin does a blocking `GET /v1/decisions` to the LAPI on
-    # *every* request. A deploy burst (runner Action API polls, `docker
-    # manifest inspect` per own image, containerd pulls, smoke probes) fires
-    # hundreds of parallel registry calls; LAPI saturation pushed the lookup
-    # past the plugin timeout, and the bouncer fail-closed with 403 - which
-    # containerd surfaces as ErrImagePull/ImagePullBackOff on the next pod.
-    # This route only serves authenticated OCI traffic (registry tokens,
-    # basic-auth already handled by gitea) and scanners get nothing useful
-    # from /v2, so there is no bruteforce surface to protect here.
     - match: Host(`gcr.forust.xyz`) && PathPrefix(`/v2`)
       kind: Rule
       services:
@@ -42,7 +29,7 @@ spec:
   entryPoints:
     - websecure
   routes:
-    - match: Host(`gitea.workstation.internal`) || Host(`gitea.gigaforust.internal`)
+    - match: (Host(`gitea.workstation.internal`) || Host(`gitea.gigaforust.internal`)) || (Host(`git.workstation.internal`) || Host(`git.gigaforust.internal`))
       kind: Rule
       services:
         - name: gitea-service

glance/k8s/glance-config.yaml

@@ -177,8 +177,8 @@ data:
                   # url: https://gitssh.forust.xyz
                 # - title: gcr.forust.xyz
                   # url: https://gcr.forust.xyz/v2/
-                - title: gitea.forust.xyz
-                  url: https://gitea.forust.xyz
+                - title: git.forust.xyz
+                  url: https://git.forust.xyz
                 - title: nextcloud.forust.xyz
                   url: https://nextcloud.forust.xyz
                 - title: mc.forust.xyz

glance/k8s/glance.yaml

@@ -20,6 +20,8 @@ spec:
   selector:
     matchLabels:
       app: glance
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -57,10 +59,10 @@ spec:
           resources:
             requests:
               cpu: "50m"
-              memory: "64Mi"
+              memory: "32Mi"
             limits:
               cpu: "200m"
-              memory: "256Mi"
+              memory: "128Mi"
       volumes:
         - name: glance-config
           configMap:

headscale/k8s/routing/ingress.yaml

@@ -23,17 +23,11 @@ spec:
           port: 8080
     - match: Host(`hs.forust.xyz`) && PathPrefix(`/admin`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: headscale-ui-external
           port: 80
     - match: Host(`hs.forust.xyz`) && PathPrefix(`/metrics`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: headscale-server-external
           port: 9090
@@ -53,8 +47,6 @@ spec:
       kind: Rule
       middlewares:
         - name: headplane-prefix
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: headplane-external
           port: 3000

homarr/.env.example

@@ -0,0 +1,4 @@
+SECRET_ENCRYPTION_KEY="REPLACE_ME"
+TZ="Europe/Bratislava"
+PUID="1000"
+PGID="1000"

homarr/compose.yaml

@@ -0,0 +1,38 @@
+services:
+  homarr:
+    container_name: homarr
+    image: ghcr.io/homarr-labs/homarr:v2.1.2
+    restart: unless-stopped
+    volumes:
+      - ./appdata:/appdata
+      - /var/run/docker.sock:/var/run/docker.sock:ro
+      - ./kubeconfig:/app/config/kubeconfig:ro
+    env_file: .env
+    ports:
+      - 80:7575
+      - 81:3000
+    environment:
+      - TZ=${TZ:-Europe/Bratislava}
+      - TURBO_TELEMETRY_DISABLED=1
+      - KUBECONFIG=/app/config/kubeconfig
+    labels:
+      - "traefik.enable=true"
+      - "traefik.http.services.homarr.loadbalancer.server.port=7575"
+
+      # Prod Router
+      - "traefik.http.routers.homarr.rule=Host(`homarr.forust.xyz`)"
+      - "traefik.http.routers.homarr.entrypoints=websecure"
+      - "traefik.http.routers.homarr.tls.certresolver=letsencrypt"
+      # Local Router
+      - "traefik.http.routers.homarr-local.rule=Host(`homarr.workstation.internal`)"
+      - "traefik.http.routers.homarr-local.entrypoints=websecure"
+      - "traefik.http.routers.homarr-local.tls=true"
+      # Dev Router
+      - "traefik.http.routers.homarr-dev.rule=Host(`homarr.gigaforust.internal`)"
+      - "traefik.http.routers.homarr-dev.entrypoints=websecure"
+      - "traefik.http.routers.homarr-dev.tls=true"
+    networks:
+      - proxy
+networks:
+  proxy:
+    external: true

homarr/k8s/certificates.yaml

@@ -0,0 +1,28 @@
+# apiVersion: cert-manager.io/v1
+# kind: Certificate
+# metadata:
+#   name: home-prod-tls
+#   namespace: homarr
+# spec:
+#   secretName: home-prod-tls
+#   dnsNames:
+#     - home.forust.xyz
+#   issuerRef:
+#     name: letsencrypt-prod
+#     kind: ClusterIssuer
+# ---
+apiVersion: cert-manager.io/v1
+kind: Certificate
+metadata:
+  name: internal-wildcard-tls
+  namespace: homarr
+spec:
+  secretName: internal-wildcard-tls
+  dnsNames:
+    - "*.workstation.internal"
+    - "*.gigaforust.internal"
+    - workstation.internal
+    - gigaforust.internal
+  issuerRef:
+    name: internal-ca
+    kind: ClusterIssuer

homarr/k8s/config.yaml

@@ -0,0 +1,9 @@
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: homarr-config
+  namespace: homarr
+data:
+  TZ: "Europe/Bratislava"
+  TURBO_TELEMETRY_DISABLED: "1"
+  ENABLE_KUBERNETES: "true"

homarr/k8s/homarr.yaml

@@ -0,0 +1,81 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: homarr-service
+  namespace: homarr
+spec:
+  selector:
+    app: homarr
+  ports:
+    - port: 7575
+      targetPort: 7575
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: homarr-deployment
+  namespace: homarr
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: homarr
+  strategy:
+    type: Recreate
+  template:
+    metadata:
+      labels:
+        app: homarr
+    spec:
+      serviceAccountName: homarr
+      containers:
+        - name: homarr
+          image: ghcr.io/homarr-labs/homarr:v2.1.2
+          envFrom:
+            - configMapRef:
+                name: homarr-config
+            - secretRef:
+                name: homarr-secrets
+          ports:
+            - containerPort: 7575
+          readinessProbe:
+            httpGet:
+              path: /
+              port: 7575
+            initialDelaySeconds: 30
+            periodSeconds: 10
+            failureThreshold: 6
+          livenessProbe:
+            httpGet:
+              path: /
+              port: 7575
+            initialDelaySeconds: 60
+            periodSeconds: 30
+            failureThreshold: 3
+          volumeMounts:
+            - name: homarr-data
+              mountPath: /appdata
+          resources:
+            requests:
+              cpu: "250m"
+              memory: "350Mi"
+            limits:
+              cpu: "500m"
+              memory: "700Mi"
+      volumes:
+        - name: homarr-data
+          persistentVolumeClaim:
+            claimName: homarr-pvc
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: homarr-pvc
+  namespace: homarr
+spec:
+  resources:
+    requests:
+      storage: 2Gi
+  volumeMode: Filesystem
+  accessModes:
+    - ReadWriteOnce

homarr/k8s/ingress.yaml

@@ -0,0 +1,33 @@
+# apiVersion: traefik.io/v1alpha1
+# kind: IngressRoute
+# metadata:
+#   name: homarr-prod
+#   namespace: homarr
+# spec:
+#   entryPoints:
+#     - websecure
+#   routes:
+#     - match: Host(`home.forust.xyz`)
+#       kind: Rule
+#       services:
+#         - name: homarr-service
+#           port: 7575
+#   tls:
+#     secretName: home-prod-tls
+# ---
+apiVersion: traefik.io/v1alpha1
+kind: IngressRoute
+metadata:
+  name: homarr-local
+  namespace: homarr
+spec:
+  entryPoints:
+    - websecure
+  routes:
+    - match: Host(`home.workstation.internal`) || Host(`home.gigaforust.internal`)
+      kind: Rule
+      services:
+        - name: homarr-service
+          port: 7575
+  tls:
+    secretName: internal-wildcard-tls

homarr/k8s/namespace.yaml

@@ -0,0 +1,4 @@
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: homarr

homarr/k8s/rbac.yaml

@@ -0,0 +1,58 @@
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: homarr
+  namespace: homarr
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRole
+metadata:
+  name: homarr-readonly
+rules:
+  - apiGroups: [""]
+    resources:
+      - pods
+      - services
+      - endpoints
+      - namespaces
+      - nodes
+      - configmaps
+      - persistentvolumeclaims
+      - events
+    verbs: ["get", "list", "watch"]
+  - apiGroups: ["apps"]
+    resources:
+      - deployments
+      - statefulsets
+      - daemonsets
+      - replicasets
+    verbs: ["get", "list", "watch"]
+  - apiGroups: ["networking.k8s.io"]
+    resources:
+      - ingresses
+    verbs: ["get", "list", "watch"]
+  - apiGroups: ["traefik.io"]
+    resources:
+      - ingressroutes
+      - ingressroutetcps
+      - ingressrouteudps
+      - middlewares
+    verbs: ["get", "list", "watch"]
+  - apiGroups: ["metrics.k8s.io"]
+    resources:
+      - pods
+      - nodes
+    verbs: ["get", "list"]
+---
+apiVersion: rbac.authorization.k8s.io/v1
+kind: ClusterRoleBinding
+metadata:
+  name: homarr-readonly
+roleRef:
+  apiGroup: rbac.authorization.k8s.io
+  kind: ClusterRole
+  name: homarr-readonly
+subjects:
+  - kind: ServiceAccount
+    name: homarr
+    namespace: homarr

homarr/k8s/secrets.yaml.example

@@ -0,0 +1,9 @@
+apiVersion: v1
+kind: Secret
+metadata:
+  name: homarr-secrets
+  namespace: homarr
+type: Opaque
+stringData:
+  # openssl rand -hex 32
+  SECRET_ENCRYPTION_KEY: "REPLACE_ME"

homepages/compose.yaml

@@ -3,7 +3,7 @@ services:
     build:
       context: .
       dockerfile: Dockerfile.forust
-    image: gcr.forust.xyz/forust/forust-homepage:latest
+    image: gcr.forust.xyz/forust/forust-homepage:prod
     pull_policy: build
     # ports:
     #   - "8085:80"
@@ -35,7 +35,7 @@ services:
     build:
       context: .
       dockerfile: Dockerfile.xdfnx
-    image: gcr.forust.xyz/forust/xdfnx-homepage:latest
+    image: gcr.forust.xyz/forust/xdfnx-homepage:prod
     pull_policy: build
     restart: unless-stopped
     # ports:

homepages/forust_files/index.html

@@ -174,7 +174,7 @@
         <h2>./projects</h2>
         <ul class="repo-list">
           <li>
-            <a href="https://gitea.forust.xyz/forust/gosleep" target="_blank">forust/gosleep</a>
+            <a href="https://git.forust.xyz/forust/gosleep" target="_blank">forust/gosleep</a>
             <span class="comment">// linux sleep timer written in rust (originally in go)</span>
           </li>
         </ul>

homepages/k8s/gateway.yaml

@@ -0,0 +1,31 @@
+# Gateway API PoC for homepages. Lives next to the TLS secrets so
+# certificateRefs stay same-namespace and no ReferenceGrant is needed.
+# Listener ports must match the Traefik entryPoints (80/443),
+# otherwise Traefik marks the listener Invalid.
+# Local .internal hosts are deliberately left on IngressRoute,
+# only prod is migrated here.
+apiVersion: gateway.networking.k8s.io/v1
+kind: Gateway
+metadata:
+  name: homepages
+  namespace: homepages
+spec:
+  gatewayClassName: traefik
+  listeners:
+    - name: http
+      protocol: HTTP
+      port: 80
+      allowedRoutes:
+        namespaces:
+          from: Same
+    - name: https
+      protocol: HTTPS
+      port: 443
+      tls:
+        mode: Terminate
+        certificateRefs:
+          - name: forust-homepage-prod-tls
+          - name: xdfnx-homepage-prod-tls
+      allowedRoutes:
+        namespaces:
+          from: Same

homepages/k8s/homepages.yaml

@@ -20,6 +20,8 @@ spec:
   selector:
     matchLabels:
       app: forust-homepage
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -39,10 +41,10 @@ spec:
             failureThreshold: 3
           resources:
             requests:
-              memory: "10Mi"
+              memory: "32Mi"
               cpu: "20m"
             limits:
-              memory: "100Mi"
+              memory: "128Mi"
               cpu: "50m"
 ---
 apiVersion: v1
@@ -67,6 +69,8 @@ spec:
   selector:
     matchLabels:
       app: xdfnx-homepage
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -86,8 +90,8 @@ spec:
             failureThreshold: 3
           resources:
             requests:
-              memory: "10Mi"
+              memory: "32Mi"
               cpu: "20m"
             limits:
-              memory: "100Mi"
+              memory: "128Mi"
               cpu: "50m"

homepages/k8s/httproute.yaml

@@ -0,0 +1,69 @@
+# PoC: homepages prod hosts via Gateway API.
+# Runs alongside k8s/ingress.yaml - delete the prod IngressRoutes only after verification.
+# There are no local .internal hosts here, they stay on the local IngressRoute.
+# No per-route security middlewares: L3 enforcement moved to the host
+# firewall bouncer, so HTTPRoutes stay clean.
+---
+apiVersion: gateway.networking.k8s.io/v1
+kind: HTTPRoute
+metadata:
+  name: homepages-http-redirect
+  namespace: homepages
+spec:
+  parentRefs:
+    - name: homepages
+      kind: Gateway
+      sectionName: http
+  hostnames:
+    - forust.xyz
+    - www.forust.xyz
+    - xdfnx.cfd
+  rules:
+    - filters:
+        - type: RequestRedirect
+          requestRedirect:
+            scheme: https
+            statusCode: 301
+---
+apiVersion: gateway.networking.k8s.io/v1
+kind: HTTPRoute
+metadata:
+  name: forust-homepage-https
+  namespace: homepages
+spec:
+  parentRefs:
+    - name: homepages
+      kind: Gateway
+      sectionName: https
+  hostnames:
+    - forust.xyz
+    - www.forust.xyz
+  rules:
+    - matches:
+        - path:
+            type: PathPrefix
+            value: /
+      backendRefs:
+        - name: forust-homepage-service
+          port: 80
+---
+apiVersion: gateway.networking.k8s.io/v1
+kind: HTTPRoute
+metadata:
+  name: xdfnx-homepage-https
+  namespace: homepages
+spec:
+  parentRefs:
+    - name: homepages
+      kind: Gateway
+      sectionName: https
+  hostnames:
+    - xdfnx.cfd
+  rules:
+    - matches:
+        - path:
+            type: PathPrefix
+            value: /
+      backendRefs:
+        - name: xdfnx-homepage-service
+          port: 80

homepages/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`forust.xyz`) || Host(`www.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       priority: 10
       services:
         - name: forust-homepage-service
@@ -49,9 +46,6 @@ spec:
   routes:
     - match: Host(`xdfnx.cfd`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: xdfnx-homepage-service
           port: 80

immich/.env.example

@@ -0,0 +1,24 @@
+# You can find documentation for all the supported env variables at https://docs.immich.app/install/environment-variables
+
+# The location where your uploaded files are stored. The k8s manifests bind
+# mount /mnt/immich/library, which is the sdc9 partition - the same place, so
+# the two deployment paths look at one library.
+UPLOAD_LOCATION=/mnt/immich/library
+
+# The location where your database files are stored. Network shares are not supported for the database
+DB_DATA_LOCATION=./postgres
+
+# To set a timezone, uncomment the next line and change Etc/UTC to a TZ identifier from this list: https://en.wikipedia.org/wiki/List_of_tz_database_time_zones#List
+# TZ=Etc/UTC
+
+# The Immich version to use. You can pin this to a specific version like "v2.1.0"
+IMMICH_VERSION=v3
+
+# Connection secret for postgres. You should change it to a random password
+# Please use only the characters `A-Za-z0-9`, without special characters or spaces
+DB_PASSWORD=postgres
+
+# The values below this line do not need to be changed
+###################################################################################
+DB_USERNAME=postgres
+DB_DATABASE_NAME=immich

immich/compose.yaml

@@ -0,0 +1,63 @@
+name: immich
+
+services:
+  immich-server:
+    container_name: immich_server
+    image: ghcr.io/immich-app/immich-server:v3
+    volumes:
+      - ${UPLOAD_LOCATION}:/data
+      - /etc/localtime:/etc/localtime:ro
+    env_file:
+      - .env
+    ports:
+      - "2283:2283"
+    depends_on:
+      - redis
+      - database
+    restart: always
+    healthcheck:
+      disable: false
+
+  immich-machine-learning:
+    container_name: immich_machine_learning
+    # For hardware acceleration, add one of -[armnn, cuda, rocm, openvino, rknn] to the image tag.
+    # Example tag: ${IMMICH_VERSION:-release}-cuda
+    image: ghcr.io/immich-app/immich-machine-learning:${IMMICH_VERSION:-release}
+    # extends: # uncomment this section for hardware acceleration - see https://docs.immich.app/features/ml-hardware-acceleration
+    #   file: hwaccel.ml.yml
+    #   service: cpu # set to one of [armnn, cuda, rocm, openvino, openvino-wsl, rknn] for accelerated inference - use the `-wsl` version for WSL2 where applicable
+    volumes:
+      - model-cache:/cache
+    env_file:
+      - .env
+    restart: always
+    healthcheck:
+      disable: false
+
+  redis:
+    container_name: immich_redis
+    image: docker.io/valkey/valkey:9@sha256:418652cfb58ef879d4978c33553735d7147016032d5aefaa14c828e611eb9dfd
+    healthcheck:
+      test: redis-cli ping | grep -q PONG || exit 1
+    restart: always
+
+  database:
+    container_name: immich_postgres
+    image: ghcr.io/immich-app/postgres:16-vectorchord0.4.3-pgvectors0.2.0@sha256:1a078b237c1d9b420b0ee59147386b4aa60d3a07a8e6a402fc84a57e41b043a4
+    environment:
+      POSTGRES_PASSWORD: ${DB_PASSWORD}
+      POSTGRES_USER: ${DB_USERNAME}
+      POSTGRES_DB: ${DB_DATABASE_NAME}
+      POSTGRES_INITDB_ARGS: "--data-checksums"
+    # Uncomment the DB_STORAGE_TYPE: 'HDD' var if your database isn't stored on SSDs
+    # DB_STORAGE_TYPE: 'HDD'
+    volumes:
+      # Do not edit the next line. If you want to change the database storage location on your system, edit the value of DB_DATA_LOCATION in the .env file
+      - ${DB_DATA_LOCATION}:/var/lib/postgresql/data
+    shm_size: 128mb
+    restart: always
+    healthcheck:
+      disable: false
+
+volumes:
+  model-cache:

immich/k8s/certificates.yaml

@@ -1,8 +1,21 @@
 apiVersion: cert-manager.io/v1
 kind: Certificate
+metadata:
+  name: immich-prod-tls
+  namespace: immich
+spec:
+  secretName: immich-prod-tls
+  dnsNames:
+    - immich.forust.xyz
+  issuerRef:
+    name: letsencrypt-prod
+    kind: ClusterIssuer
+---
+apiVersion: cert-manager.io/v1
+kind: Certificate
 metadata:
   name: internal-wildcard-tls
-  namespace: userbot
+  namespace: immich
 spec:
   secretName: internal-wildcard-tls
   dnsNames:

immich/k8s/config.yaml

@@ -0,0 +1,24 @@
+apiVersion: v1
+kind: ConfigMap
+metadata:
+  name: immich-config
+  namespace: immich
+data:
+  TZ: "Europe/Bratislava"
+
+  # The database in this namespace, not the shared one in the database
+  # namespace: v3 needs VectorChord, and only the dedicated image carries it.
+  DB_HOSTNAME: "immich-postgres"
+  DB_PORT: "5432"
+  DB_USERNAME: "immich"
+  DB_DATABASE_NAME: "immich"
+  DB_SSL_MODE: "disable"
+  DB_VECTOR_EXTENSION: "vectorchord"
+
+  REDIS_HOSTNAME: "immich-valkey"
+  REDIS_PORT: "6379"
+
+  # Traefik is the only client of the server, and it is a pod: the address immich
+  # sees is inside the node's pod CIDR. Without this the server does not trust
+  # X-Forwarded-For and every request looks like it came from Traefik itself.
+  IMMICH_TRUSTED_PROXIES: "10.244.0.0/24"

immich/k8s/immich.yaml

@@ -0,0 +1,95 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: immich-service
+  namespace: immich
+spec:
+  selector:
+    app: immich
+  ports:
+    - name: http
+      port: 2283
+      targetPort: 2283
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: immich-deployment
+  namespace: immich
+  labels:
+    app: immich
+spec:
+  replicas: 2
+  selector:
+    matchLabels:
+      app: immich
+  template:
+    metadata:
+      labels:
+        app: immich
+    spec:
+      containers:
+        - name: immich
+          image: ghcr.io/immich-app/immich-server:v3
+          envFrom:
+            - configMapRef:
+                name: immich-config
+            - secretRef:
+                name: immich-secrets
+          ports:
+            - name: http
+              containerPort: 2283
+          volumeMounts:
+            - name: immich-data
+              mountPath: /data
+          # The first boot runs migrations and warms the transcoder, which can
+          # take minutes, so liveness has to wait on the startup probe.
+          startupProbe:
+            httpGet:
+              path: /api/server/ping
+              port: http
+            failureThreshold: 60
+            periodSeconds: 10
+            timeoutSeconds: 5
+          readinessProbe:
+            httpGet:
+              path: /api/server/ping
+              port: http
+            periodSeconds: 10
+            timeoutSeconds: 5
+          livenessProbe:
+            httpGet:
+              path: /api/server/ping
+              port: http
+            initialDelaySeconds: 30
+            periodSeconds: 30
+            timeoutSeconds: 5
+          # Only the request is scheduled against, and the node is already
+          # oversubscribed (5.58 of 6 cores requested) while actually running
+          # at about 1.5. So the request states what this sits at while idle -
+          # tens of millicores - and the limit leaves room for the burst that
+          # matters: thumbnails, transcodes and metadata extraction.
+          #
+          # The limit used to be 2Gi, but the server OOMKilled on boot while
+          # chewing through a backlog of unprocessed assets (API +Workers in
+          # one container spike well past idle).
+          resources:
+            requests:
+              cpu: "100m"
+              memory: "512Mi"
+            limits:
+              cpu: "1500m"
+              memory: "4Gi"
+      volumes:
+        - name: immich-data
+          # The library lives on the node's own disk, not in a PVC. A PVC here
+          # meant declaring a size up front for data that does not exist yet,
+          # on a provisioner that cannot grow it, and the only copy of the
+          # photos was one `kubectl delete namespace` away.
+          #
+          # Directory, not DirectoryOrCreate, on purpose: if sdc9 is not
+          # mounted, this must fail loudly instead of quietly writing the
+          # library onto the root filesystem.
+          hostPath:
+            path: /mnt/immich/library
+            type: Directory

immich/k8s/ingress.yaml

@@ -0,0 +1,33 @@
+apiVersion: traefik.io/v1alpha1
+kind: IngressRoute
+metadata:
+  name: immich-prod
+  namespace: immich
+spec:
+  entryPoints:
+    - websecure
+  routes:
+    - match: Host(`immich.forust.xyz`)
+      kind: Rule
+      services:
+        - name: immich-service
+          port: 2283
+  tls:
+    secretName: immich-prod-tls
+---
+apiVersion: traefik.io/v1alpha1
+kind: IngressRoute
+metadata:
+  name: immich-local
+  namespace: immich
+spec:
+  entryPoints:
+    - websecure
+  routes:
+    - match: Host(`immich.workstation.internal`) || Host(`immich.gigaforust.internal`)
+      kind: Rule
+      services:
+        - name: immich-service
+          port: 2283
+  tls:
+    secretName: internal-wildcard-tls

immich/k8s/machine-learning.yaml

@@ -0,0 +1,94 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: immich-machine-learning
+  namespace: immich
+spec:
+  selector:
+    app: immich-machine-learning
+  ports:
+    - name: http
+      port: 3003
+      targetPort: 3003
+---
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: immich-machine-learning-deployment
+  namespace: immich
+  labels:
+    app: immich-machine-learning
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: immich-machine-learning
+  strategy:
+    type: Recreate
+  template:
+    metadata:
+      labels:
+        app: immich-machine-learning
+    spec:
+      containers:
+        - name: immich-machine-learning
+          image: ghcr.io/immich-app/immich-machine-learning:v3
+          envFrom:
+            - configMapRef:
+                name: immich-config
+            - secretRef:
+                name: immich-secrets
+          ports:
+            - name: http
+              containerPort: 3003
+          volumeMounts:
+            - name: model-cache
+              mountPath: /cache
+          # The first request pulls a model over the internet, so a cold start
+          # is slower than a container start.
+          startupProbe:
+            httpGet:
+              path: /ping
+              port: http
+            failureThreshold: 60
+            periodSeconds: 5
+            timeoutSeconds: 5
+          readinessProbe:
+            httpGet:
+              path: /ping
+              port: http
+            periodSeconds: 10
+            timeoutSeconds: 5
+          livenessProbe:
+            httpGet:
+              path: /ping
+              port: http
+            initialDelaySeconds: 30
+            periodSeconds: 30
+            timeoutSeconds: 5
+          # Same reasoning as the server: the request covers the idle cost
+          # only, because the node has no spare cores to schedule against.
+          # Recognition is the burst - a busy import wants both cores.
+          resources:
+            requests:
+              cpu: "100m"
+              memory: "1Gi"
+            limits:
+              cpu: "2000m"
+              memory: "3Gi"
+      volumes:
+        - name: model-cache
+          persistentVolumeClaim:
+            claimName: immich-model-cache-pvc
+---
+apiVersion: v1
+kind: PersistentVolumeClaim
+metadata:
+  name: immich-model-cache-pvc
+  namespace: immich
+spec:
+  accessModes:
+    - ReadWriteOnce
+  resources:
+    requests:
+      storage: 2Gi

immich/k8s/namespace.yaml

@@ -0,0 +1,5 @@
+# yaml-language-server: $schema=kubernetes
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: immich

immich/k8s/postgres.yaml

@@ -0,0 +1,145 @@
+# Immich's own database, separate from the shared postgres in the database
+# namespace. It has to be separate: v3 checks the VectorChord version at startup
+# and refuses to boot without it, VectorChord needs its .so in
+# shared_preload_libraries, and that can only be read when postmaster starts.
+# So the shared instance would have to be rebuilt on a custom image carrying
+# vchord and restarted - for every consumer of it (authentik, gitea, netbox,
+# netronome, penpot, statuspage). Not worth it for one photo library.
+apiVersion: v1
+kind: Service
+metadata:
+  name: immich-postgres
+  namespace: immich
+  labels:
+    app: immich-postgres
+spec:
+  selector:
+    app: immich-postgres
+  ports:
+    - name: postgres
+      port: 5432
+      targetPort: postgres
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: immich-postgres
+  namespace: immich
+  labels:
+    app: immich-postgres
+spec:
+  serviceName: immich-postgres
+  replicas: 1
+  selector:
+    matchLabels:
+      app: immich-postgres
+  template:
+    metadata:
+      labels:
+        app: immich-postgres
+    spec:
+      containers:
+        - name: postgres
+          # v3.x expects vchord for its vector work and vectors (pgvecto.rs)
+          # for some index types. This image ships both and preloads them, plus
+          # its own shared_buffers and wal settings, through
+          # /etc/postgresql/postgresql.conf - which its entrypoint reaches via
+          # `postgres -c config_file=...` in the image CMD.
+          #
+          # So there is deliberately no `command:` here. Overriding it replaces
+          # that config_file, and it also loses the step where the entrypoint
+          # drops from root to the postgres user: postmaster then starts as
+          # root and refuses to run.
+          image: ghcr.io/immich-app/postgres:16-vectorchord0.4.3-pgvectors0.2.0@sha256:1a078b237c1d9b420b0ee59147386b4aa60d3a07a8e6a402fc84a57e41b043a4
+          env:
+            - name: POSTGRES_USER
+              value: immich
+            - name: POSTGRES_DB
+              value: immich
+            - name: POSTGRES_PASSWORD
+              valueFrom:
+                secretKeyRef:
+                  name: immich-secrets
+                  key: DB_PASSWORD
+            # Only read when the data directory is empty, so the checksums are
+            # decided here and never again.
+            - name: POSTGRES_INITDB_ARGS
+              value: --data-checksums
+            # The postgres-data volume lives on sdc, which is rotational. The
+            # SSD template is the default; HDD only changes the planner costs
+            # (effective_io_concurrency, random_page_cost), nothing structural.
+            - name: DB_STORAGE_TYPE
+              value: HDD
+            - name: TZ
+              valueFrom:
+                configMapKeyRef:
+                  name: immich-config
+                  key: TZ
+          ports:
+            - name: postgres
+              containerPort: 5432
+          volumeMounts:
+            - name: postgres-data
+              mountPath: /var/lib/postgresql/data
+            # The upstream compose file asks docker for 128mb of shm. Kubernetes
+            # gives every container 64mb, which is not what postmaster expects
+            # for parallel query workers and the WAL writer.
+            - name: shm
+              mountPath: /dev/shm
+          # Probes use a generous timeout on purpose: the data lives on a
+          # rotational disk on a loaded single node, and pg_isready can take
+          # seconds during WAL recovery. A 1s timeout kills the container
+          # mid-recovery and restarts the spiral.
+          #
+          # Budgets are sized for HDD stalls, not for a healthy disk: fsync of
+          # a single file was observed taking 70s under node IO pressure, so
+          # the startup budget is 15 minutes and liveness tolerates 5 minutes
+          # of unresponsiveness. Killing a stalled-but-healthy postmaster only
+          # buys another full WAL replay, which is more IO, not less.
+          startupProbe:
+            exec:
+              command: ["sh", "-c", "pg_isready -U immich -d immich"]
+            failureThreshold: 180
+            periodSeconds: 5
+            timeoutSeconds: 5
+          readinessProbe:
+            exec:
+              command: ["sh", "-c", "pg_isready -U immich -d immich"]
+            periodSeconds: 10
+            timeoutSeconds: 5
+          livenessProbe:
+            exec:
+              command: ["sh", "-c", "pg_isready -U immich -d immich"]
+            initialDelaySeconds: 30
+            periodSeconds: 60
+            timeoutSeconds: 10
+            failureThreshold: 5
+          # The image template sets shared_buffers to 512MB, and the vchord and
+          # vectors workers are Rust binaries with a real RSS footprint on top
+          # of postmaster, checkpointer and friends. 1Gi was enough to start
+          # the server but the vectors worker kept dying in it, so the limit
+          # sits at 2Gi. The request stays at the idle cost.
+          resources:
+            requests:
+              cpu: "50m"
+              memory: "256Mi"
+            limits:
+              cpu: "1000m"
+              memory: "2Gi"
+      volumes:
+        - name: shm
+          emptyDir:
+            medium: Memory
+            sizeLimit: 128Mi
+  volumeClaimTemplates:
+    - metadata:
+        name: postgres-data
+      spec:
+        accessModes: ["ReadWriteOnce"]
+        # Retain: this is the metadata for a library that only exists in one
+        # place, and local-path cannot expand a bound volume, so this size has
+        # to hold until the library is rebuilt or dumped elsewhere.
+        storageClassName: local-path-retain
+        resources:
+          requests:
+            storage: 32Gi

immich/k8s/secrets.yaml.example

@@ -0,0 +1,13 @@
+apiVersion: v1
+kind: Secret
+metadata:
+  name: immich-secrets
+  namespace: immich
+type: Opaque
+stringData:
+  # Creates the immich superuser in this namespace's own postgres on first
+  # boot, and is the same value the server connects with. Nothing outside the
+  # immich namespace needs it. Letters and digits only: immich reads this into
+  # a connection string.
+  DB_PASSWORD: "changeme"
+  REDIS_PASSWORD: "changeme"

immich/k8s/valkey.yaml

@@ -0,0 +1,84 @@
+apiVersion: v1
+kind: Service
+metadata:
+  name: immich-valkey
+  namespace: immich
+  labels:
+    app: immich-valkey
+spec:
+  clusterIP: None
+  selector:
+    app: immich-valkey
+  ports:
+    - name: valkey
+      port: 6379
+      targetPort: valkey
+---
+apiVersion: apps/v1
+kind: StatefulSet
+metadata:
+  name: immich-valkey
+  namespace: immich
+  labels:
+    app: immich-valkey
+spec:
+  serviceName: immich-valkey
+  replicas: 1
+  selector:
+    matchLabels:
+      app: immich-valkey
+  template:
+    metadata:
+      labels:
+        app: immich-valkey
+    spec:
+      containers:
+        - name: valkey
+          image: docker.io/valkey/valkey:9.1.2-alpine
+          command:
+            - sh
+            - -c
+            - valkey-server --appendonly yes --save 30 1 --loglevel warning --requirepass "$REDIS_PASSWORD"
+          envFrom:
+            - secretRef:
+                name: immich-secrets
+          ports:
+            - name: valkey
+              containerPort: 6379
+          volumeMounts:
+            - name: valkey-data
+              mountPath: /data
+          # Same reasoning as postgres: 1s probe timeouts flap on a loaded
+          # single node with rotational storage.
+          startupProbe:
+            exec:
+              command: ["sh", "-c", 'valkey-cli --pass "$REDIS_PASSWORD" ping | grep -q PONG']
+            failureThreshold: 20
+            periodSeconds: 5
+            timeoutSeconds: 5
+          readinessProbe:
+            exec:
+              command: ["sh", "-c", 'valkey-cli --pass "$REDIS_PASSWORD" ping | grep -q PONG']
+            periodSeconds: 10
+            timeoutSeconds: 5
+          livenessProbe:
+            exec:
+              command: ["sh", "-c", 'valkey-cli --pass "$REDIS_PASSWORD" ping | grep -q PONG']
+            initialDelaySeconds: 20
+            periodSeconds: 20
+            timeoutSeconds: 5
+          resources:
+            requests:
+              cpu: "25m"
+              memory: "64Mi"
+            limits:
+              cpu: "250m"
+              memory: "256Mi"
+  volumeClaimTemplates:
+    - metadata:
+        name: valkey-data
+      spec:
+        accessModes: ["ReadWriteOnce"]
+        resources:
+          requests:
+            storage: 1Gi

kener/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`status.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: kener-service
           port: 3000

kener/k8s/kener.yaml

@@ -20,6 +20,8 @@ spec:
   selector:
     matchLabels:
       app: kener
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:

loki/k8s/alloy-values.yaml

@@ -7,18 +7,32 @@
 
 controller:
   type: daemonset
+
+# config-reloader sidecar: p95 33M, max 43M. The chart keeps it at the top level,
+# not under `alloy:`.
+configReloader:
   resources:
     requests:
-      memory: "128Mi"
-      cpu: "50m"
+      memory: "32Mi"
+      cpu: "10m"
     limits:
-      memory: "512Mi"
-      cpu: "500m"
+      memory: "128Mi"
 
 image:
   tag: "v1.19.2"
 
 alloy:
+  # p95 275M, max 287M. Alloy tails every pod log and ships it to Loki, so it sits
+  # on the same IronWolf read path the node is I/O bound on. Request is set at p95.
+  # The chart key is `alloy.resources`. `controller.resources` is ignored silently,
+  # which is why this pod shipped with no limits at all.
+  resources:
+    requests:
+      memory: "288Mi"
+      cpu: "50m"
+    limits:
+      memory: "512Mi"
+
   configMap:
     create: true
     content: |

loki/k8s/loki-values.yaml

@@ -39,6 +39,16 @@ loki:
       local:
         directory: /var/loki/rules
 
+# p95 84M, max 85M for the rules sidecar that shares the singleBinary pod.
+# The chart exposes it as `sidecar.resources`, shared with any other sidecar.
+sidecar:
+  resources:
+    requests:
+      memory: "96Mi"
+      cpu: "10m"
+    limits:
+      memory: "192Mi"
+
 singleBinary:
   replicas: 1
   persistence:
@@ -47,10 +57,10 @@ singleBinary:
     storageClass: local-path-retain
   resources:
     requests:
-      memory: "512Mi"
+      memory: "256Mi"
       cpu: "200m"
     limits:
-      memory: "2Gi"
+      memory: "1Gi"
       cpu: "1000m"
 
 # Zeroed: unused in SingleBinary mode (chart validation requires it).
@@ -63,12 +73,25 @@ backend:
 
 gateway:
   replicas: 1
+  # Single node: chart default is required podAntiAffinity on hostname +
+  # RollingUpdate 25%/25% (effective maxUnavailable=0 at replicas=1).
+  # That deadlocks the rollout: the new pod stays Unschedulable while the
+  # old one lives, and the old one never leaves while the new one is not
+  # Ready. Null clears the default (an empty map would deep-merge with it
+  # and keep the required rule); maxUnavailable=1 allows a brief gateway
+  # outage during rollouts instead of a stuck deploy.
+  affinity: null
+  deploymentStrategy:
+    type: RollingUpdate
+    rollingUpdate:
+      maxSurge: 1
+      maxUnavailable: 1
   resources:
     requests:
-      memory: "64Mi"
+      memory: "32Mi"
       cpu: "50m"
     limits:
-      memory: "256Mi"
+      memory: "128Mi"
       cpu: "300m"
 
 monitoring:

metube/compose.yaml

@@ -1,6 +1,6 @@
 services:
   metube:
-    image: ghcr.io/alexta69/metube:2026.09.27
+    image: ghcr.io/alexta69/metube:2026.09.29
     container_name: metube
     restart: unless-stopped
     # ports:

metube/k8s/metube.yaml

@@ -20,6 +20,8 @@ spec:
   selector:
     matchLabels:
       app: metube
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -27,7 +29,7 @@ spec:
     spec:
       containers:
         - name: metube
-          image: ghcr.io/alexta69/metube:2026.09.27
+          image: ghcr.io/alexta69/metube:2026.09.29
           envFrom:
             - configMapRef:
                 name: metube-config
@@ -36,12 +38,13 @@ spec:
           volumeMounts:
             - name: downloads
               mountPath: /downloads
+          # p95 72M, max 80M over 7 days. Was 600Mi/2Gi.
           resources:
             requests:
-              memory: "600Mi"
+              memory: "96Mi"
               cpu: "400m"
             limits:
-              memory: "2Gi"
+              memory: "384Mi"
               cpu: "1700m"
       volumes:
         - name: downloads

n8n/compose.yaml

@@ -1,6 +1,6 @@
 services:
   n8n:
-    image: docker.n8n.io/n8nio/n8n:2.41.3
+    image: docker.n8n.io/n8nio/n8n:2.42.3
     container_name: n8n
     restart: unless-stopped
     environment:

n8n/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`n8n.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: n8n-service
           port: 5678

n8n/k8s/n8n.yaml

@@ -20,6 +20,8 @@ spec:
   selector:
     matchLabels:
       app: n8n
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -27,7 +29,7 @@ spec:
     spec:
       containers:
         - name: n8n
-          image: docker.n8n.io/n8nio/n8n:2.41.3
+          image: docker.n8n.io/n8nio/n8n:2.42.3
           envFrom:
             - configMapRef:
                 name: n8n-config

netbird/compose.yaml

@@ -2,7 +2,7 @@ name: netbird
 
 services:
   netbird-server:
-    image: netbirdio/netbird-server:0.79.0
+    image: netbirdio/netbird-server:0.80.0
     container_name: netbird-server
     restart: unless-stopped
     environment:
@@ -87,7 +87,7 @@ services:
       - proxy
 
   dashboard:
-    image: netbirdio/dashboard:v2.93.0
+    image: netbirdio/dashboard:v2.94.0
     container_name: netbird-dashboard
     restart: unless-stopped
     environment:

netbird/k8s/ingress.yaml

@@ -7,12 +7,18 @@ spec:
   entryPoints:
     - websecure
   routes:
+    # NO crowdsec-bouncer on the API routes. These are the mesh client's own
+    # endpoints: gRPC-gateway management calls plus signal/relay long-polling,
+    # authenticated by NetBird's token rather than by a login form. A ban here
+    # is self-defeating - the client needs the mesh to reach anything else, so
+    # CrowdSec banning it locks the peer out of the network it needs to
+    # function. It also backfires: a banned peer keeps retrying, every retry
+    # is another 403, and LePresidente/http-generic-403-bf turns five 403s in
+    # ten seconds into a 4h ban, so one 403 loop kept re-arming the ban.
+    # netbird-local below has always been exempt; this makes prod match.
     - match: Host(`nb.forust.xyz`) && (PathPrefix(`/signalexchange.SignalExchange/`) || PathPrefix(`/management.ManagementService/`) || PathPrefix(`/management.ProxyService/`))
       kind: Rule
       priority: 100
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: netbird-server-service
           port: 80
@@ -20,18 +26,12 @@ spec:
     - match: Host(`nb.forust.xyz`) && (PathPrefix(`/relay`) || PathPrefix(`/ws-proxy/`) || PathPrefix(`/api`) || PathPrefix(`/oauth2`))
       kind: Rule
       priority: 100
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: netbird-server-service
           port: 80
     - match: Host(`nb.forust.xyz`)
       kind: Rule
       priority: 1
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: netbird-dashboard-service
           port: 80

netbird/k8s/netbird.yaml

@@ -39,6 +39,8 @@ spec:
   selector:
     matchLabels:
       app: netbird-server
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -46,7 +48,7 @@ spec:
     spec:
       containers:
         - name: netbird-server
-          image: netbirdio/netbird-server:0.79.0
+          image: netbirdio/netbird-server:0.80.0
           command: ["/bin/sh", "/opt/netbird/entrypoint.sh", "--config", "/run/netbird/config.yaml"]
           envFrom:
             - configMapRef:
@@ -88,12 +90,13 @@ spec:
             periodSeconds: 30
             timeoutSeconds: 5
             failureThreshold: 5
+          # p95 97M, max 102M over 7 days. Was 256Mi/1Gi.
           resources:
             requests:
-              memory: "256Mi"
-              cpu: "250m"
+              memory: "128Mi"
+              cpu: "100m"
             limits:
-              memory: "1Gi"
+              memory: "384Mi"
               cpu: "1000m"
       volumes:
         - name: netbird-data
@@ -130,6 +133,8 @@ spec:
   selector:
     matchLabels:
       app: netbird-dashboard
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -137,7 +142,7 @@ spec:
     spec:
       containers:
         - name: dashboard
-          image: netbirdio/dashboard:v2.93.0
+          image: netbirdio/dashboard:v2.94.0
           envFrom:
             - configMapRef:
                 name: netbird-config
@@ -162,10 +167,10 @@ spec:
             failureThreshold: 5
           resources:
             requests:
-              memory: "64Mi"
+              memory: "32Mi"
               cpu: "50m"
             limits:
-              memory: "256Mi"
+              memory: "128Mi"
               cpu: "300m"
 ---
 apiVersion: v1

netbox/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`netbox.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: netbox-service
           port: 8080

netbox/k8s/netbox.yaml

@@ -164,7 +164,7 @@ spec:
               memory: "256Mi"
             limits:
               cpu: "1"
-              memory: "1Gi"
+              memory: "512Mi"
       volumes:
         - name: netbox-config
           configMap:

netronome/compose.yaml

@@ -1,6 +1,6 @@
 services:
   netronome:
-    image: ghcr.io/autobrr/netronome:v0.14.1
+    image: ghcr.io/autobrr/netronome:v0.15.0
     restart: unless-stopped
     container_name: netronome
     ports:

netronome/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`nm.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: netronome-service
           port: 7575

netronome/k8s/netronome.yaml

@@ -23,6 +23,8 @@ spec:
   selector:
     matchLabels:
       app: netronome
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:
@@ -30,7 +32,7 @@ spec:
     spec:
       containers:
         - name: netronome
-          image: ghcr.io/autobrr/netronome:v0.14.1
+          image: ghcr.io/autobrr/netronome:v0.15.0
           ports:
             - name: netronome-port
               protocol: TCP
@@ -51,8 +53,8 @@ spec:
                   key: NETRONOME__DB_PASSWORD
           resources:
             requests:
-              memory: "100Mi"
+              memory: "64Mi"
               cpu: "100m"
             limits:
-              memory: "512Mi"
+              memory: "256Mi"
               cpu: "500m"

nextcloud/k8s/routing/ingress.yaml

@@ -12,8 +12,6 @@ spec:
       kind: Rule
       middlewares:
         - name: nextcloud-chain@file
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: nextcloud-apache
           port: 11000
@@ -32,8 +30,6 @@ spec:
     - match: Host(`nextcloud.workstation.internal`) || Host(`nextcloud.gigaforust.internal`)
       kind: Rule
       middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
         - name: nextcloud-chain@file
       services:
         - name: nextcloud-apache

portainer/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`portainer.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: portainer-service
           port: 9000

portainer/k8s/portainer.yaml

@@ -20,6 +20,8 @@ spec:
   selector:
     matchLabels:
       app: portainer
+  strategy:
+    type: Recreate
   template:
     metadata:
       labels:

postgres/k8s/postgres.yaml

@@ -60,16 +60,26 @@ spec:
               command: ["pg_isready", "-U", "postgres", "-d", "postgres"]
             initialDelaySeconds: 10
             periodSeconds: 10
+            # Generous timeout: on an I/O-bound single node even exec can take
+            # seconds, and a 1s default kills a healthy postgres mid-recovery.
+            timeoutSeconds: 5
           startupProbe:
             exec:
               command: ["pg_isready", "-U", "postgres", "-d", "postgres"]
-            failureThreshold: 30
+            # Crash recovery on an I/O-starved single node can fsync for 10+
+            # minutes; killing postgres mid-recovery restarts the fsync from
+            # zero and loops forever. 90x10s = 15 minutes of grace.
+            failureThreshold: 90
             periodSeconds: 10
           livenessProbe:
             exec:
               command: ["pg_isready", "-U", "postgres", "-d", "postgres"]
             initialDelaySeconds: 30
             periodSeconds: 20
+            # Same I/O reasoning as readiness, plus more misses before a kill:
+            # restarting postgres on a loaded node only makes recovery longer.
+            timeoutSeconds: 5
+            failureThreshold: 5
           resources:
             requests:
               memory: "512Mi"

prometheus-stack/compose.yaml

@@ -1,6 +1,6 @@
 services:
   grafana:
-    image: grafana/grafana:13.2.2
+    image: grafana/grafana:13.2.3
     container_name: prometheus-grafana
     restart: unless-stopped
     env_file:

prometheus-stack/k8s/alerts.yaml

@@ -27,6 +27,33 @@ spec:
             summary: "Pod is crash looping"
             description: "Container {{ $labels.container }} in {{ $labels.namespace }}/{{ $labels.pod }} is in CrashLoopBackOff."
 
+        - alert: ContainerOOMKilled
+          expr: max_over_time(kube_pod_container_status_terminated_reason{reason="OOMKilled"}[15m]) >= 1
+          for: 5m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Container was OOMKilled"
+            description: "Container {{ $labels.container }} in {{ $labels.namespace }}/{{ $labels.pod }} was killed for exceeding its memory limit. Raise the limit or reduce the workload."
+
+        - alert: ContainerRestartingTooOften
+          expr: max by (namespace, pod, container) (increase(kube_pod_container_status_restarts_total[30m])) > 3
+          for: 5m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Container restarting too often"
+            description: "Container {{ $labels.container }} in {{ $labels.namespace }}/{{ $labels.pod }} restarted {{ $value }} times in the last 30 minutes."
+
+        - alert: PodEvicted
+          expr: max_over_time(kube_pod_status_reason{reason="Evicted"}[15m]) >= 1
+          for: 5m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Pod was evicted"
+            description: "Pod {{ $labels.namespace }}/{{ $labels.pod }} was evicted, usually for node disk or memory pressure."
+
         - alert: PersistentVolumeClaimFillingUp
           expr: kubelet_volume_stats_available_bytes / kubelet_volume_stats_capacity_bytes < 0.15
           for: 15m

prometheus-stack/k8s/crowdsec-alerts.yaml

@@ -0,0 +1,41 @@
+apiVersion: monitoring.coreos.com/v1
+kind: PrometheusRule
+metadata:
+  name: crowdsec
+  namespace: prometheus
+  labels:
+    release: prometheus-stack
+spec:
+  groups:
+    - name: crowdsec
+      rules:
+        - alert: CrowdsecFirewallBouncerStale
+          expr: |
+            absent(cs_lapi_bouncer_requests_total{bouncer="firewall-workstation"})
+            or sum(rate(cs_lapi_bouncer_requests_total{bouncer="firewall-workstation"}[10m])) == 0
+          for: 15m
+          labels:
+            severity: critical
+          annotations:
+            summary: "Firewall bouncer stopped pulling decisions"
+            description: "No LAPI pulls from firewall-workstation for 15 minutes. L3 enforcement is decaying: existing bans expire, new ones never land. Check the systemd unit on the node."
+
+        - alert: CrowdsecDecisionsSpike
+          expr: |
+            sum(cs_active_decisions) - sum(cs_active_decisions offset 30m) > 20
+          for: 5m
+          labels:
+            severity: warning
+          annotations:
+            summary: "Spike in active CrowdSec decisions"
+            description: "Active decisions grew by more than 20 in 30 minutes (current: {{ $value }}). Possible ban storm or self-ban - check cscli decisions list."
+
+        - alert: CrowdsecLAPIDecisionErrors
+          expr: |
+            sum(rate(cs_lapi_decisions_ko_total[5m])) > 0
+          for: 10m
+          labels:
+            severity: warning
+          annotations:
+            summary: "CrowdSec LAPI decision errors"
+            description: "LAPI is failing decision lookups. Bouncers may be failing open. Check LAPI logs and DB health."

prometheus-stack/k8s/grafana-values.yaml

@@ -26,6 +26,25 @@ grafana:
   service:
     port: 80
 
+  # p95 391M, observed max 1046M with no limit at all. Request is set at p95 so the
+  # scheduler sees reality; the limit is a manual exception above the 1.3x max
+  # formula, because a single query burst reached 1046M.
+  resources:
+    requests:
+      memory: "416Mi"
+      cpu: 100m
+    limits:
+      memory: "1536Mi"
+
+  # One block covers both the dashboards and datasources sidecars (p95 91M / 80M).
+  sidecar:
+    resources:
+      requests:
+        memory: "96Mi"
+        cpu: 10m
+      limits:
+        memory: "192Mi"
+
   additionalDataSources:
     - name: Loki
       type: loki
@@ -48,13 +67,21 @@ prometheus:
               storage: 40Gi
     resources:
       requests:
-        memory: "700Mi"
+        memory: "768Mi"
         cpu: 200m
       limits:
-        memory: "2Gi"
+        memory: "2560Mi"
 alertmanager:
   alertmanagerSpec:
     configSecret: alertmanager-config
+    # p95 66M, max 68M. Silences and notification state live here, so the request
+    # stays above p95 to keep the pod out of the eviction candidates.
+    resources:
+      requests:
+        memory: "96Mi"
+        cpu: 10m
+      limits:
+        memory: "192Mi"
     storage:
       volumeClaimTemplate:
         spec:
@@ -66,6 +93,39 @@ alertmanager:
             requests:
               storage: 20Gi
 
+# p95 103M, max 107M, and it grows with the number of cluster objects.
+kube-state-metrics:
+  # Values key is the dependency name from Chart.yaml, not the `kubeStateMetrics`
+  # condition key. Setting resources under `kubeStateMetrics:` is silently ignored.
+  resources:
+    requests:
+      memory: "128Mi"
+      cpu: 50m
+    limits:
+      memory: "256Mi"
+
+# p95 39M, max 40M. One per node, so it scales with node count.
+prometheus-node-exporter:
+  resources:
+    requests:
+      memory: "32Mi"
+      cpu: 20m
+    limits:
+      memory: "128Mi"
+
+# p95 75M, max 75M. Creates and reconciles every PrometheusRule in the cluster.
+prometheusOperator:
+  resources:
+    requests:
+      memory: "96Mi"
+      cpu: 50m
+    limits:
+      memory: "192Mi"
+
+# config-reloader sidecars (p95 33M, max 43M) are not covered: the chart does not
+# template `prometheusSpec.configReloader.resources`, so there is no values key
+# for them. They keep shipping with requests only.
+
 defaultRules:
   disabled:
     CPUThrottlingHigh: true

prometheus-stack/k8s/ingress.yaml

@@ -9,9 +9,6 @@ spec:
   routes:
     - match: Host(`grafana.forust.xyz`)
       kind: Rule
-      middlewares:
-        - name: crowdsec-bouncer
-          namespace: crowdsec
       services:
         - name: prometheus-stack-grafana
           port: 80