Health checks, retries, and auto-rollback

MonkeysCloud verifies every deploy actually works — and reverts automatically if it doesn’t.

Updated Aug 3, 2026Deployments & CI/CD

Deploying isn’t done when the build finishes — it’s done when your app is actually serving traffic correctly. Health checks make that automatic.

How it works

After a deploy, MonkeysCloud pings your app’s health endpoint until it responds successfully.

deploy:
  health_check:
    path: /health
    interval: 10s
    timeout: 5s
    retries: 3

Retries

If the first check fails, MonkeysCloud retries for a while — apps often need a few seconds to warm up. Only when the retries are exhausted is the deploy treated as failed.

Auto-rollback

If the health check never passes, or the app’s error rate climbs past your threshold, MonkeysCloud automatically rolls back to the previous good deployment. Your users keep working, and you get a notification with the details.

Recommended practice

  • Expose a /health endpoint that checks your database connection, not just that the web server is running.
  • Keep the interval and timeout modest (seconds, not minutes) so rollbacks are fast.
  • Set an alert on rollback events so the whole team learns what happened.
Health checks, retries, and auto-rollback