.github/workflows/deploy-vps.yml

name: Deploy to VPS

on:
  workflow_run:
    workflows: ["Publish to Docker Hub"]
    types: [completed]
  workflow_dispatch:

permissions:
  contents: read

jobs:
  deploy:
    if: >-
      (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success')
      && vars.DEPLOY_ENABLED == 'true'
    name: Deploy OmniRoute to VPS
    runs-on: ubuntu-latest
    steps:
      - name: Check VPS SSH reachability from runner
        id: reach
        env:
          # Pass the host via env (never interpolate a secret straight into the
          # script body) so /dev/tcp gets a shell variable, not inlined text.
          VPS_HOST: ${{ secrets.VPS_HOST }}
        run: |
          set -uo pipefail
          # A GitHub-hosted runner can only deploy when it can actually open a TCP
          # connection to the VPS SSH port. The Local VPS lives on a private LAN and
          # the Akamai host firewalls :22 to known IPs, so the runner is routinely
          # unable to reach it (`dial tcp ***:22: i/o timeout`). Treat "unreachable
          # from the runner" as a SKIP — the real deploys are run manually from an
          # allowed network via the deploy-vps-local / deploy-vps-akamai skills — so
          # an unreachable host no longer red-fails every release/push pipeline.
          # When the host IS reachable, the deploy step below still runs in full and
          # its health gate surfaces any genuine deploy failure.
          if timeout 15 bash -c 'exec 3<>"/dev/tcp/${VPS_HOST}/22"' 2>/dev/null; then
            echo "reachable=true" >> "$GITHUB_OUTPUT"
            echo "✅ VPS_HOST:22 reachable from the runner — proceeding with deploy."
          else
            echo "reachable=false" >> "$GITHUB_OUTPUT"
            echo "::warning title=Auto-deploy skipped::VPS_HOST:22 is not reachable from this GitHub runner (private LAN / firewalled). Deploy manually with the deploy-vps-local or deploy-vps-akamai skill."
          fi

      - name: Deploy via SSH
        if: steps.reach.outputs.reachable == 'true'
        uses: appleboy/ssh-action@v1
        with:
          host: ${{ secrets.VPS_HOST }}
          username: ${{ secrets.VPS_USER }}
          key: ${{ secrets.VPS_SSH_KEY }}
          port: 22
          timeout: 60s
          command_timeout: 15m
          script: |
            set -euo pipefail

            echo "=== Updating OmniRoute ==="
            npm install -g omniroute@latest
            INSTALLED_VERSION=$(omniroute --version 2>/dev/null | tr -d '[:space:]' || echo "unknown")
            echo "Installed CLI version: $INSTALLED_VERSION"

            # Recreate the PM2 process instead of `pm2 restart`. A bare restart
            # re-runs whatever script path was saved earlier; after the build-output
            # reorg (app/ -> dist/, .next -> .build/next) a process pinned to the old
            # app/server-ws.mjs path can no longer start, and the node process dies
            # while PM2 still reports "online" — so the box never binds :20128.
            # Always launch via the `omniroute` bin so .env is loaded and the dist/
            # layout is resolved correctly.
            echo "=== (Re)creating PM2 process via bin ==="
            pm2 delete omniroute 2>/dev/null || true
            pm2 start omniroute --name omniroute -- --port 20128
            pm2 save

            # Health gate: fail the deploy unless the box actually reports healthy.
            # Poll /api/monitoring/health for "status":"healthy" (a deeper signal than
            # a static page 200 — it confirms the app booted, not just that a port is
            # bound). Boot can take a while after a native-module/build-layout change,
            # so poll up to ~3min before giving up.
            echo "=== Health Check (gates the deploy) ==="
            ok=0
            for i in $(seq 1 36); do
              BODY=$(curl -sf -m 5 http://localhost:20128/api/monitoring/health 2>/dev/null || true)
              if printf '%s' "$BODY" | grep -q '"status":"healthy"'; then
                ok=1
                echo "✅ /api/monitoring/health -> healthy (attempt $i) — version $INSTALLED_VERSION"
                break
              fi
              echo "… not healthy yet (attempt $i/36), retrying in 5s"
              sleep 5
            done
            if [ "$ok" != "1" ]; then
              echo "❌ Health check failed — /api/monitoring/health never reported healthy after ~3min"
              echo "--- recent PM2 logs ---"
              pm2 logs omniroute --lines 40 --nostream || true
              exit 1
            fi
            echo "=== Deploy complete ==="