ci: harden deploy + add health monitor (#5)

* ci(deploy): retry rsync on Hostinger runner-IP throttle + best-effort inventory guard

Pilot surfaced an exit-255 SSH connection timeout (documented Hostinger
rate-limiting of concurrent runner IPs). rsync now retries 4x/25s; the plugin
count guard is best-effort so a throttled check can't fail a good deploy.

Agent-Session: cto harden-deploy

* ci: add daily Deploy Health Monitor (drift + deploy-failure)

Scheduled GitHub Actions check (runs on GitHub infra, not the dead local cron):
for each app repo, compares the latest production deployment SHA to main and
flags failed deploy runs. Opens one de-duplicated 'deploy-health' issue when
something is off — the alarm that was missing when deploys silently broke.

Agent-Session: cto deploy-health-monitor

---------

Co-authored-by: peptiderepo <peptiderepo@users.noreply.github.com>

.github/workflows/deploy-app.yml

@@ -1,19 +1,19 @@
 name: Reusable Deploy (staging → smoke → production)
 
 # Single shared deploy pipeline for every peptiderepo plugin and the theme.
-# Callers pass `target_path` (the wp-content subpath) and `secrets: inherit`,
+# Callers pass `target_path` + `secrets: inherit`; deploy logic lives HERE so a
-# and the deploy logic lives HERE — fix it once instead of in six copy-pasted
+# fix lands once instead of in six copy-pasted deploy.yml files.
-# per-repo deploy.yml files (the drift that broke every deploy in May 2026).
 #
-# Staging deploy → smoke (informational) → production deploy (env-gated) with a
+# Hardened against the documented Hostinger behaviour of rate-limiting concurrent
-# pre/post active-plugin-count guard (the 2026-04-23 WP-core-clobber tripwire),
+# GitHub-runner IPs (connection timeouts, exit 255): the rsync is retried with
-# LiteSpeed purge, and a health check.
+# backoff, and the plugin-inventory guard is best-effort so a throttled count
+# check never fails an otherwise-good deploy.
 
 on:
   workflow_call:
     inputs:
       target_path:
-        description: "wp-content subpath, e.g. wp-content/plugins/prautoblogger or wp-content/themes/peptide-starter"
+        description: "wp-content subpath, e.g. wp-content/plugins/prautoblogger"
         required: true
         type: string
       run_smoke:
@@ -22,18 +22,12 @@ on:
         default: true
         type: boolean
     secrets:
-      SSH_PRIVATE_KEY:
+      SSH_PRIVATE_KEY: {required: true}
-        required: true
+      SSH_HOST: {required: true}
-      SSH_HOST:
+      SSH_PORT: {required: true}
-        required: true
+      SSH_USERNAME: {required: true}
-      SSH_PORT:
+      STAGING_WP_ADMIN_USER: {required: false}
-        required: true
+      STAGING_WP_ADMIN_PASSWORD: {required: false}
-      SSH_USERNAME:
-        required: true
-      STAGING_WP_ADMIN_USER:
-        required: false
-      STAGING_WP_ADMIN_PASSWORD:
-        required: false
 
 permissions:
   contents: read
@@ -45,21 +39,27 @@ jobs:
     environment: staging
     steps:
       - uses: actions/checkout@v4
-      - name: Deploy via rsync → staging
+      - name: Deploy via rsync → staging (retry-on-throttle)
         run: |
           mkdir -p ~/.ssh
           echo "${{ secrets.SSH_PRIVATE_KEY }}" > ~/.ssh/deploy_key
           chmod 600 ~/.ssh/deploy_key
           ssh-keyscan -p "${{ secrets.SSH_PORT }}" "${{ secrets.SSH_HOST }}" >> ~/.ssh/known_hosts 2>/dev/null || true
           BASE="/home/${{ secrets.SSH_USERNAME }}/domains/peptiderepo.com/public_html"
+          SSHOPT="-i ~/.ssh/deploy_key -p ${{ secrets.SSH_PORT }} -o StrictHostKeyChecking=accept-new -o ConnectTimeout=30 -o BatchMode=yes"
+          n=0
+          until [ "$n" -ge 4 ]; do
             rsync -avz --delete \
               --exclude '.git' --exclude '.github' --exclude 'tests' --exclude 'node_modules' \
               --exclude 'vendor' --exclude 'composer.json' --exclude 'composer.lock' \
               --exclude 'phpunit.xml' --exclude 'phpunit.xml.dist' --exclude 'phpcs.xml.dist' \
               --exclude '.phpcs.xml.dist' --exclude 'ARCHITECTURE.md' --exclude 'CONVENTIONS.md' \
               --exclude 'CHANGELOG.md' --exclude 'qa-reviews' \
-            -e "ssh -i ~/.ssh/deploy_key -p ${{ secrets.SSH_PORT }} -o StrictHostKeyChecking=accept-new -o ConnectTimeout=30 -o BatchMode=yes" \
+              -e "ssh $SSHOPT" \
-            ./ "${{ secrets.SSH_USERNAME }}@${{ secrets.SSH_HOST }}:$BASE/staging/${{ inputs.target_path }}/"
+              ./ "${{ secrets.SSH_USERNAME }}@${{ secrets.SSH_HOST }}:$BASE/staging/${{ inputs.target_path }}/" && break
+            n=$((n+1)); echo "staging rsync failed (attempt $n/4) — Hostinger may be throttling runner IPs; retrying in 25s..."; sleep 25
+          done
+          [ "$n" -lt 4 ] || { echo "::error::staging rsync failed after 4 attempts"; exit 1; }
       - name: Staging health check
         run: |
           sleep 5
@@ -79,42 +79,44 @@ jobs:
 
   deploy-production:
     name: Deploy to Production
-    needs: [deploy-staging]   # smoke is informational; not a prod gate
+    needs: [deploy-staging]
     runs-on: ubuntu-latest
     environment: production
     steps:
       - uses: actions/checkout@v4
-      - name: Deploy via rsync → production (plugin-inventory guard)
+      - name: Deploy via rsync → production (retry + inventory guard)
         run: |
           mkdir -p ~/.ssh
           echo "${{ secrets.SSH_PRIVATE_KEY }}" > ~/.ssh/deploy_key
           chmod 600 ~/.ssh/deploy_key
           ssh-keyscan -p "${{ secrets.SSH_PORT }}" "${{ secrets.SSH_HOST }}" >> ~/.ssh/known_hosts 2>/dev/null || true
           BASE="/home/${{ secrets.SSH_USERNAME }}/domains/peptiderepo.com/public_html"
-          SSHC="ssh -i ~/.ssh/deploy_key -p ${{ secrets.SSH_PORT }} -o StrictHostKeyChecking=accept-new -o ConnectTimeout=30 -o BatchMode=yes ${{ secrets.SSH_USERNAME }}@${{ secrets.SSH_HOST }}"
+          SSHOPT="-i ~/.ssh/deploy_key -p ${{ secrets.SSH_PORT }} -o StrictHostKeyChecking=accept-new -o ConnectTimeout=30 -o BatchMode=yes"
-          PRE=$($SSHC "wp --path=$BASE plugin list --status=active --field=name 2>/dev/null | wc -l" || echo 0)
+          PRE=$(ssh $SSHOPT "${{ secrets.SSH_USERNAME }}@${{ secrets.SSH_HOST }}" "wp --path=$BASE plugin list --status=active --field=name 2>/dev/null | wc -l" 2>/dev/null || echo 0)
-          echo "Pre-deploy active plugins: $PRE"
+          echo "Pre-deploy active plugins (best-effort): $PRE"
+          n=0
+          until [ "$n" -ge 4 ]; do
             rsync -avz --delete \
               --exclude '.git' --exclude '.github' --exclude 'tests' --exclude 'node_modules' \
               --exclude 'vendor' --exclude 'composer.json' --exclude 'composer.lock' \
               --exclude 'phpunit.xml' --exclude 'phpunit.xml.dist' --exclude 'phpcs.xml.dist' \
               --exclude '.phpcs.xml.dist' --exclude 'ARCHITECTURE.md' --exclude 'CONVENTIONS.md' \
               --exclude 'CHANGELOG.md' --exclude 'qa-reviews' \
-            -e "ssh -i ~/.ssh/deploy_key -p ${{ secrets.SSH_PORT }} -o StrictHostKeyChecking=accept-new -o ConnectTimeout=30 -o BatchMode=yes" \
+              -e "ssh $SSHOPT" \
-            ./ "${{ secrets.SSH_USERNAME }}@${{ secrets.SSH_HOST }}:$BASE/${{ inputs.target_path }}/"
+              ./ "${{ secrets.SSH_USERNAME }}@${{ secrets.SSH_HOST }}:$BASE/${{ inputs.target_path }}/" && break
-          POST=$($SSHC "wp --path=$BASE plugin list --status=active --field=name 2>/dev/null | wc -l" || echo 0)
+            n=$((n+1)); echo "prod rsync failed (attempt $n/4) — retrying in 25s..."; sleep 25
-          echo "Post-deploy active plugins: $POST"
+          done
-          if [ "${PRE:-0}" -gt 0 ] && [ "${POST:-0}" -lt "${PRE:-0}" ]; then
+          [ "$n" -lt 4 ] || { echo "::error::production rsync failed after 4 attempts"; exit 1; }
-            echo "ERROR: active plugin count dropped ${PRE} -> ${POST} after deploy — aborting as failed."
+          POST=$(ssh $SSHOPT "${{ secrets.SSH_USERNAME }}@${{ secrets.SSH_HOST }}" "wp --path=$BASE plugin list --status=active --field=name 2>/dev/null | wc -l; find wp-content/litespeed/ -type f -delete 2>/dev/null; touch .lscache_purge 2>/dev/null || true" 2>/dev/null | head -1 || echo 0)
-            exit 1
+          echo "Post-deploy active plugins (best-effort): $POST"
+          if [ "${PRE:-0}" -gt 0 ] && [ "${POST:-0}" -gt 0 ] && [ "${POST:-0}" -lt "${PRE:-0}" ]; then
+            echo "::error::active plugin count dropped ${PRE} -> ${POST} after deploy"; exit 1
           fi
-          $SSHC "cd $BASE && find wp-content/litespeed/ -type f -delete 2>/dev/null; touch .lscache_purge 2>/dev/null || true"
-          echo "LiteSpeed cache purged."
       - name: Production health check
         run: |
           sleep 5
           S=$(curl -s -o /dev/null -w '%{http_code}' https://peptiderepo.com)
-          if [ "$S" != "200" ] && [ "$S" != "403" ]; then echo "ERROR: site returned HTTP $S"; exit 1; fi
+          if [ "$S" != "200" ] && [ "$S" != "403" ]; then echo "::error::site returned HTTP $S"; exit 1; fi
           echo "Production healthy (HTTP $S)"
       - name: Cleanup
         if: always()

.github/workflows/deploy-health.yml

@@ -0,0 +1,70 @@
+name: Deploy Health Monitor
+
+# Daily drift + deploy-failure check across all peptiderepo app repos.
+# Runs on GitHub's scheduler (independent of the local board cron that died).
+# Opens ONE de-duplicated 'deploy-health' issue here when something is off.
+
+on:
+  schedule:
+    - cron: '0 13 * * *'
+  workflow_dispatch:
+
+permissions:
+  contents: read
+  issues: write
+  actions: read
+
+jobs:
+  check:
+    name: Check prod-vs-main drift + last deploy status
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/github-script@v7
+        with:
+          script: |
+            const owner = 'peptiderepo';
+            const repos = [
+              'prautoblogger','peptide-repo-core','peptide-search-ai',
+              'peptide-news-plugin','peptide-reconstitution-calculator','peptide-starter-theme'
+            ];
+            const short = s => s ? String(s).substring(0,7) : '—';
+            const problems = [];
+            for (const repo of repos) {
+              try {
+                const main = (await github.rest.repos.getBranch({owner, repo, branch:'main'})).data.commit.sha;
+                let runConc = 'none', runSha = '';
+                try {
+                  const runs = (await github.rest.actions.listWorkflowRuns({owner, repo, workflow_id:'deploy.yml', per_page:1})).data.workflow_runs;
+                  if (runs.length) { runConc = runs[0].conclusion || runs[0].status; runSha = runs[0].head_sha; }
+                } catch (e) { runConc = 'no-deploy-workflow'; }
+                let prodSha = null, prodState = 'none';
+                const deps = (await github.rest.repos.listDeployments({owner, repo, environment:'production', per_page:1})).data;
+                if (deps.length) {
+                  prodSha = deps[0].sha;
+                  const sts = (await github.rest.repos.listDeploymentStatuses({owner, repo, deployment_id: deps[0].id, per_page:1})).data;
+                  if (sts.length) prodState = sts[0].state;
+                }
+                if (runConc === 'failure') problems.push(`**${repo}**: latest deploy run failed (\`${short(runSha)}\`)`);
+                if (prodSha && prodSha !== main) problems.push(`**${repo}**: prod behind main — prod=\`${short(prodSha)}\` vs main=\`${short(main)}\` (undeployed commits)`);
+                if (prodSha && prodState !== 'success') problems.push(`**${repo}**: last production deployment state is \`${prodState}\``);
+                core.info(`${repo}: main=${short(main)} lastRun=${runConc} prodDeploy=${short(prodSha)}/${prodState}`);
+              } catch (e) {
+                problems.push(`**${repo}**: monitor error — ${e.message}`);
+              }
+            }
+            const summary = problems.length
+              ? `### ⚠️ Deploy health: ${problems.length} issue(s)\n\n` + problems.map(p=>`- ${p}`).join('\n')
+              : `### ✅ All app repos current; last deploy healthy`;
+            await core.summary.addRaw(summary).write();
+            if (problems.length) {
+              const existing = (await github.rest.issues.listForRepo({owner, repo: context.repo.repo, labels:'deploy-health', state:'open'})).data;
+              if (existing.length === 0) {
+                await github.rest.issues.create({ owner, repo: context.repo.repo,
+                  title: `Deploy health: ${problems.length} issue(s) detected`,
+                  body: summary + `\n\n_Opened by deploy-health.yml. Close once resolved; the monitor won't reopen while this stays open._`,
+                  labels: ['deploy-health'] });
+                core.warning(`Opened deploy-health issue (${problems.length} problems)`);
+              } else {
+                core.warning(`${problems.length} problems; deploy-health issue already open (#${existing[0].number})`);
+              }
+            }