aca-schedule.yml

name: ACA Manual Scale (disabled schedule — ADR-018)

# ADR-018: INF-STG-EU_EHDS runs 24/7. Scheduled triggers disabled.
# The start/stop jobs remain available via workflow_dispatch for manual
# maintenance windows (e.g., cost investigations, planned downtime).
on:
  # schedule:
  #   - cron: "0 5 * * 1-5"   # disabled — was start (07:00 Europe/Berlin)
  #   - cron: "0 18 * * *"    # disabled — was stop  (20:00 Europe/Berlin)
  workflow_dispatch:
    inputs:
      action:
        description: "Action to run"
        type: choice
        options: [start, stop]
        required: true

permissions:
  id-token: write
  contents: read

concurrency:
  group: "azure-schedule"
  cancel-in-progress: false

env:
  RESOURCE_GROUP: rg-mvhd-dev
  PG_SERVER: pg-mvhd-dev

jobs:
  determine-action:
    name: Determine Action
    runs-on: ubuntu-latest
    timeout-minutes: 2
    outputs:
      action: ${{ steps.decide.outputs.action }}
    steps:
      - id: decide
        run: |
          if [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
            echo "action=${{ inputs.action }}" >> "$GITHUB_OUTPUT"
          elif [ "${{ github.event.schedule }}" = "0 5 * * 1-5" ]; then
            echo "action=start" >> "$GITHUB_OUTPUT"
          else
            echo "action=stop" >> "$GITHUB_OUTPUT"
          fi

  stop:
    name: Scale Down (Stop)
    needs: determine-action
    if: needs.determine-action.outputs.action == 'stop'
    runs-on: ubuntu-latest
    timeout-minutes: 20
    steps:
      - name: Azure Login (OIDC)
        uses: azure/login@v2
        with:
          client-id: ${{ secrets.AZURE_CLIENT_ID }}
          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}

      - name: Scale all Container Apps to 0
        run: |
          echo "=== Scaling 13 Container Apps to 0 replicas ==="
          APPS=(
            mvhd-neo4j mvhd-nats mvhd-keycloak mvhd-vault
            mvhd-neo4j-proxy mvhd-ui mvhd-controlplane
            mvhd-dp-fhir mvhd-dp-omop mvhd-identityhub mvhd-issuerservice
            mvhd-tenant-mgr mvhd-provision-mgr
          )
          for app in "${APPS[@]}"; do
            echo "  $app -> 0/0"
            az containerapp update \
              --name "$app" --resource-group "$RESOURCE_GROUP" \
              --min-replicas 0 --max-replicas 0 \
              -o none || echo "::warning::Failed to scale $app"
          done

      - name: Stop PostgreSQL Flexible Server
        run: |
          echo "=== Stopping $PG_SERVER ==="
          az postgres flexible-server stop \
            --name "$PG_SERVER" --resource-group "$RESOURCE_GROUP" \
            -o none || echo "::warning::PG stop failed (may already be stopped)"

      - name: Stop summary
        if: always()
        run: |
          echo "## Off-Hours Scale-Down" >> "$GITHUB_STEP_SUMMARY"
          echo "" >> "$GITHUB_STEP_SUMMARY"
          echo "**Time:** $(date -u +%Y-%m-%dT%H:%M:%SZ)" >> "$GITHUB_STEP_SUMMARY"
          echo "**Status:** All 13 Container Apps scaled to 0, PostgreSQL stopped" >> "$GITHUB_STEP_SUMMARY"
          echo "**Next start:** Monday-Friday 05:00 UTC (07:00 Europe/Berlin CEST)" >> "$GITHUB_STEP_SUMMARY"

  start:
    name: Scale Up (Start)
    needs: determine-action
    if: needs.determine-action.outputs.action == 'start'
    runs-on: ubuntu-latest
    timeout-minutes: 30
    steps:
      - name: Azure Login (OIDC)
        uses: azure/login@v2
        with:
          client-id: ${{ secrets.AZURE_CLIENT_ID }}
          tenant-id: ${{ secrets.AZURE_TENANT_ID }}
          subscription-id: ${{ secrets.AZURE_SUBSCRIPTION_ID }}

      - name: Start PostgreSQL Flexible Server
        run: |
          echo "=== Starting $PG_SERVER ==="
          az postgres flexible-server start \
            --name "$PG_SERVER" --resource-group "$RESOURCE_GROUP" \
            -o none || echo "::warning::PG start failed (may already be running)"

      - name: Wait for PostgreSQL to be Ready
        run: |
          for i in $(seq 1 30); do
            state=$(az postgres flexible-server show \
              --name "$PG_SERVER" --resource-group "$RESOURCE_GROUP" \
              --query "state" -o tsv 2>/dev/null || echo "Unknown")
            echo "  attempt $i: state=$state"
            if [ "$state" = "Ready" ]; then
              echo "PostgreSQL is Ready"
              break
            fi
            sleep 10
          done

      - name: Restore Container App scale profiles
        run: |
          echo "=== Restoring Container App replica profiles ==="
          # Apps with min=1 max=1 (singleton stateful / small services)
          for app in mvhd-neo4j mvhd-nats mvhd-keycloak mvhd-vault mvhd-identityhub mvhd-issuerservice mvhd-tenant-mgr mvhd-provision-mgr; do
            echo "  $app -> 1/1"
            az containerapp update \
              --name "$app" --resource-group "$RESOURCE_GROUP" \
              --min-replicas 1 --max-replicas 1 \
              -o none || echo "::warning::Failed to scale $app"
          done
          # Apps with min=1 max=2
          for app in mvhd-neo4j-proxy mvhd-controlplane mvhd-dp-fhir mvhd-dp-omop; do
            echo "  $app -> 1/2"
            az containerapp update \
              --name "$app" --resource-group "$RESOURCE_GROUP" \
              --min-replicas 1 --max-replicas 2 \
              -o none || echo "::warning::Failed to scale $app"
          done
          # UI: min=1 max=3
          echo "  mvhd-ui -> 1/3"
          az containerapp update \
            --name mvhd-ui --resource-group "$RESOURCE_GROUP" \
            --min-replicas 1 --max-replicas 3 \
            -o none || echo "::warning::Failed to scale mvhd-ui"

      - name: Wait for stateful services to be reachable
        run: |
          echo "Waiting 75s for Neo4j + Vault pods to start..."
          sleep 75

      - name: Re-bootstrap Vault (in-memory gotcha #1)
        run: |
          echo "=== Re-running bootstrap job (Vault secrets + keys) ==="
          az containerapp job start \
            --name mvhd-vault-bootstrap --resource-group "$RESOURCE_GROUP" \
            -o none || echo "::warning::Vault bootstrap job start failed"

      # Neo4j is now persistent on Azure Files (ADR-017). The seed job is no
      # longer required on every start — data survives the 0/0 → 1/1 cycle.
      # If the database is empty (fresh storage account) the one-time bootstrap
      # workflow `bootstrap-stateful.yml` should be run manually instead.

      - name: Verify UI is reachable
        id: verify
        run: |
          UI_FQDN=$(az containerapp show --name mvhd-ui \
            --resource-group "$RESOURCE_GROUP" \
            --query "properties.configuration.ingress.fqdn" -o tsv)
          echo "ui_url=https://${UI_FQDN}" >> "$GITHUB_OUTPUT"
          echo "Checking https://${UI_FQDN}..."
          for i in $(seq 1 30); do
            code=$(curl -sf -o /dev/null -w "%{http_code}" "https://${UI_FQDN}" 2>/dev/null || echo "000")
            if [ "$code" = "200" ]; then
              echo "UI is healthy (HTTP 200) after $((i * 10))s"
              exit 0
            fi
            echo "  attempt $i: HTTP $code"
            sleep 10
          done
          echo "::warning::UI did not return 200 within 300s — may need manual check"

      - name: Start summary
        if: always()
        run: |
          echo "## On-Hours Scale-Up" >> "$GITHUB_STEP_SUMMARY"
          echo "" >> "$GITHUB_STEP_SUMMARY"
          echo "**Time:** $(date -u +%Y-%m-%dT%H:%M:%SZ)" >> "$GITHUB_STEP_SUMMARY"
          echo "**Target:** ${{ steps.verify.outputs.ui_url }}" >> "$GITHUB_STEP_SUMMARY"
          echo "**Status:** PostgreSQL started, 13 Container Apps scaled up, Vault re-bootstrapped (Neo4j data persists via Azure Files — see ADR-017)" >> "$GITHUB_STEP_SUMMARY"
          echo "**Next stop:** Today 18:00 UTC (20:00 Europe/Berlin CEST)" >> "$GITHUB_STEP_SUMMARY"