Why a Render service keeps restarting, and how to hear about it
A Render service keeps restarting because its process exits and Render starts it again. The logs show the exit code before each restart, and code 137 means the instance ran out of memory. Many restarts within a few minutes is a crash loop. Restarts hours apart is a service that falls over under load. For the general picture, see why a container keeps restarting.
Where Render shows it
- The service’s Events feed lists each failure and restart.
- The logs show the exit code and the last lines before it.
- The Metrics tab shows memory: a flat line at the limit means the instance is too small, a steady climb a leak.
Common causes on Render
- Out of memory. Exit code 137. See out of memory on Render.
- A failing health check. The path in Settings, under Health Check, does not return a 2xx or 3xx status. The app may listen on the wrong port: it must use the
PORTvariable, on0.0.0.0. - A crash on boot. A missing environment variable or a failed migration. During a deploy the same error repeats until the deploy times out and fails.
What Render tells you
Render notifies by email or Slack when a deploy fails, when a running service becomes unhealthy, and when it suspends a service that kept failing to start for 24 hours. A service that restarts and comes back healthy each time sends nothing, and the Events feed is per service.