improve health checks across all nomad job specs

- traefik: TCP → HTTP check on /ping (enable ping entrypoint)
- gitea: check path → /api/healthz
- jellyfin: TCP → HTTP check on /health
- glance: TCP → HTTP check on /
- sonarr/prowlarr: check path / → /ping (×2 checks each)
- ntfy/transfer/deluge/openreader/authelia/pgadmin: add name and port to existing checks
- postgres: remove invalid TCP check (Connect-enabled service)
- unifi: TCP → script check via curl (macvlan host isolation workaround)

.gitea/workflows/ci.yml

@@ -52,3 +52,39 @@ jobs:
             echo "==> $f"
             nomad job validate "$f"
           done
+
+  image-pull:
+    name: Docker image pull validation
+    runs-on: ubuntu-latest
+    # Only run on PRs that touch nomad job specs
+    if: github.event_name == 'pull_request'
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Pull changed images
+        run: |
+          git fetch origin ${{ github.base_ref }}
+          IMAGES=$(git diff origin/${{ github.base_ref }}...HEAD -- '*.nomad.hcl' \
+            | grep '^+\s*image\s*=' \
+            | grep -oP '"[^"]+:[^"]+"' \
+            | tr -d '"' \
+            | sort -u || true)
+
+          if [ -z "$IMAGES" ]; then
+            echo "No image changes detected, skipping pull."
+            exit 0
+          fi
+
+          FAILED=0
+          while IFS= read -r image; do
+            echo "==> Pulling $image"
+            if ! docker pull "$image"; then
+              echo "ERROR: Failed to pull $image"
+              FAILED=1
+            fi
+          done <<< "$IMAGES"
+
+          exit $FAILED

2-nomad-config/1-data/pgadmin.nomad.hcl

@@ -58,8 +58,10 @@ job "pgadmin" {
         ]
 
         check {
+          name     = "alive"
           type     = "http"
           path     = "/"
+          port     = "http"
           interval = "10s"
           timeout  = "2s"
         }

2-nomad-config/1-data/postgres.nomad.hcl

@@ -7,6 +7,10 @@ job "postgres" {
       connect {
         sidecar_service {}
       }
+
+      # Note: TCP checks are not valid for Connect-enabled services (runs through
+      # Envoy sidecar). Postgres is a single-writer DB that we never canary, so
+      # observable health via Consul is lower priority than other services.
     }
 
     task "postgres" {

2-nomad-config/2-ingress/authelia.nomad.hcl

@@ -33,8 +33,10 @@ job "authelia" {
       }
 
       check {
+        name     = "alive"
         type     = "http"
         path     = "/health"
+        port     = "http"
         interval = "10s"
         timeout  = "2s"
       }

2-nomad-config/2-ingress/traefik.nomad.hcl

@@ -35,7 +35,8 @@ job "traefik" {
 
       check {
         name     = "alive"
-        type     = "tcp"
+        type     = "http"
+        path     = "/ping"
         port     = "api"
         interval = "10s"
         timeout  = "2s"
@@ -85,6 +86,9 @@ api:
   dashboard: true
   insecure: true
 
+ping:
+  entryPoint: traefik
+
 providers:
   file:
     directory: "/etc/traefik/configs/"

2-nomad-config/act-runner.nomad.hcl

@@ -29,7 +29,6 @@ job "act-runner" {
 
       env = {
         GITEA_INSTANCE_URL = "https://gitea-1ef0bea6b75a4fd3e9393a9f7f7e4b02.othrayte.one"
-        CONFIG_FILE        = "/secrets/runner-config.yml"
       }
 
       # Required SOPS key:
@@ -43,19 +42,6 @@ EOF
         env         = true
       }
 
-      # Limit which images/labels the runner will accept so it doesn't pick up
-      # unrelated workloads if more runners are added later.
-      template {
-        data        = <<EOF
-runner:
-  labels:
-    - "ubuntu-latest:docker://node:20-bookworm"
-    - "ubuntu-22.04:docker://node:20-bookworm"
-    - "ubuntu-24.04:docker://node:20-bookworm"
-EOF
-        destination = "secrets/runner-config.yml"
-      }
-
       resources {
         cpu        = 200
         memory     = 256

2-nomad-config/deluge.nomad.hcl

@@ -54,8 +54,10 @@ EOH
       ]
 
       check {
+        name     = "alive"
         type     = "http"
         path     = "/"
+        port     = "http"
         interval = "10s"
         timeout  = "2s"
       }
@@ -78,8 +80,10 @@ EOH
       }
 
       check {
+        name     = "alive"
         type     = "http"
         path     = "/"
+        port     = "http"
         interval = "10s"
         timeout  = "2s"
       }

2-nomad-config/gitea.nomad.hcl

@@ -32,8 +32,10 @@ job "gitea" {
       ]
 
       check {
+        name     = "alive"
         type     = "http"
-        path     = "/"
+        path     = "/api/healthz"
+        port     = "http"
         interval = "10s"
         timeout  = "2s"
       }

2-nomad-config/glance.nomad.hcl

@@ -12,7 +12,7 @@ job "glance" {
       driver = "docker"
 
       config {
-        image = "glanceapp/glance:latest"
+        image = "glanceapp/glance:v0.7.8"
         ports = ["http"]
         volumes = [
           "local/glance.yml:/app/config/glance.yml",
@@ -30,7 +30,8 @@ job "glance" {
 
         check {
           name     = "alive"
-          type     = "tcp"
+          type     = "http"
+          path     = "/"
           port     = "http"
           interval = "10s"
           timeout  = "2s"

2-nomad-config/jellyfin.nomad.hcl

@@ -28,10 +28,11 @@ job "jellyfin" {
 
         check {
           name     = "alive"
-          type     = "tcp"
+          type     = "http"
+          path     = "/health"
           port     = "http"
           interval = "10s"
-          timeout  = "2s"
+          timeout  = "5s"
         }
       }
 

2-nomad-config/ntfy.nomad.hcl

@@ -33,8 +33,10 @@ job "ntfy" {
       ]
 
       check {
+        name     = "alive"
         type     = "http"
         path     = "/healthz"
+        port     = "http"
         interval = "10s"
         timeout  = "2s"
       }

2-nomad-config/openreader.nomad.hcl

@@ -31,8 +31,10 @@ job "openreader" {
       ]
 
       check {
+        name     = "alive"
         type     = "http"
         path     = "/"
+        port     = "http"
         interval = "10s"
         timeout  = "2s"
       }
@@ -48,8 +50,10 @@ job "openreader" {
       }
 
       check {
+        name     = "alive"
         type     = "http"
         path     = "/"
+        port     = "http"
         interval = "10s"
         timeout  = "2s"
       }

2-nomad-config/prowlarr.nomad.hcl

@@ -34,8 +34,9 @@ job "prowlarr" {
       ]
 
       check {
+        name     = "alive"
         type     = "http"
-        path     = "/"
+        path     = "/ping"
         interval = "10s"
         timeout  = "2s"
       }
@@ -51,8 +52,9 @@ job "prowlarr" {
       }
 
       check {
+        name     = "alive"
         type     = "http"
-        path     = "/"
+        path     = "/ping"
         interval = "10s"
         timeout  = "2s"
       }

2-nomad-config/sonarr.nomad.hcl

@@ -37,8 +37,9 @@ job "sonarr" {
       ]
 
       check {
+        name     = "alive"
         type     = "http"
-        path     = "/"
+        path     = "/ping"
         interval = "10s"
         timeout  = "2s"
       }
@@ -54,8 +55,9 @@ job "sonarr" {
       }
 
       check {
+        name     = "alive"
         type     = "http"
-        path     = "/"
+        path     = "/ping"
         interval = "10s"
         timeout  = "2s"
       }

2-nomad-config/transfer.nomad.hcl

@@ -16,8 +16,10 @@ job "transfer" {
       ]
 
       check {
+        name     = "alive"
         type     = "http"
         path     = "/"
+        port     = "http"
         interval = "10s"
         timeout  = "2s"
       }

2-nomad-config/unifi.nomad.hcl

@@ -21,6 +21,26 @@ job "unifi-network" {
         UNIFI_STDOUT       = "true"
       }
 
+      # Register in Consul so Traefik and health checks can find it.
+      # address_mode=driver uses the macvlan IP (192.168.1.50) rather than the host IP.
+      service {
+        name         = "unifi-network"
+        port         = 8443
+        address_mode = "driver"
+
+        # TCP/HTTP checks from the Consul agent can't reach the macvlan IP (host↔macvlan
+        # isolation). Use a script check instead — it runs inside the container via
+        # docker exec and connects to localhost:8443 directly.
+        check {
+          name     = "alive"
+          type     = "script"
+          command  = "/usr/bin/curl"
+          args     = ["-sk", "--max-time", "5", "-o", "/dev/null", "https://localhost:8443"]
+          interval = "30s"
+          timeout  = "10s"
+        }
+      }
+
       volume_mount {
         volume      = "unraid_appdata_unifi_network"
         destination = "/unifi" # Expected root directory (contains data, log, cert subdirs)

cicd-plan.md

@@ -188,8 +188,8 @@ Most jobs already have Consul health checks — these can use `health_check = "c
 | frigate    | ✅            | ✅ `single-node-writer` | ⚠️ same — rolling                                                                 |
 | glance     | ✅            | no                      | ✅ yes                                                                            |
 | transfer   | ✅            | ✅ `single-node-writer` | ⚠️ rolling                                                                        |
-| openreader | ❌            | ✅ `single-node-writer` | ⚠️ add check first, then rolling                                                  |
-| unifi      | ❌            | ✅ `single-node-writer` | ⚠️ add check first, then rolling                                                  |
+| openreader | ✅ `/`        | ✅ `single-node-writer` | ⚠️ rolling                                                                        |
+| unifi      | ✅ script     | ✅ `single-node-writer` | ⚠️ rolling                                                                        |
 | traefik    | (ingress)     | ✅                      | ⚠️ rolling — downtime risk, promote quickly                                       |
 | authelia   | (ingress)     | ✅                      | ✅ stateless config, canary fine                                                  |
 | renovate   | batch job     | n/a                     | n/a — no deployment model                                                         |
@@ -296,10 +296,10 @@ exit 1
 - [x] **Phase 1a**: Create `act-runner.nomad.hcl` + Terraform wrapper, register runner token in Gitea, get a hello-world workflow green
 - [x] **Phase 1b**: Add `terraform fmt` + `terraform validate -backend=false` workflow — no secrets needed
 - [x] **Phase 1c**: Add Nomad validate step — add `NOMAD_ADDR` + read-only `NOMAD_TOKEN` to Gitea secrets
-- [ ] **Phase 2**: Add image pull validation step to the workflow
+- [x] **Phase 2**: Add image pull validation step to the workflow
 - [ ] **Phase 3a**: Add `update` stanzas to ntfy and glance (simplest, no volume conflict)
-- [ ] **Phase 3b**: Add rolling `update` stanzas to remaining service jobs (jellyfin, sonarr, etc.)
-- [ ] **Phase 3c**: Add health checks to openreader and unifi before adding update stanzas
+- [ ] **Phase 3b**: Add rolling `update` stanzas to remaining service jobs (jellyfin, sonarr, prowlarr, deluge, gitea, immich, transfer, frigate, openreader, unifi, authelia, traefik)
+- [x] **Phase 3c**: Add health checks to openreader and unifi before adding update stanzas
 - [ ] **Phase 4a**: Add on-push workflow that runs `terraform apply -auto-approve` using full credential set
 - [ ] **Phase 4b**: Add deployment promotion/revert polling script
 - [ ] **Phase 4c**: Wire ntfy notifications for promote/revert outcomes