App Runner fails to update on any environment variable change

Question

I have an App Runner service successfully running using a VPC connector. It is using port 3000 and running an HTTP health check on that port on the endpoint `/health`. I can see those requests in the logs. Everything seems fine.

When I add a new environment variable via the console, the update fails after the health check (it does not say that the health check **failed**, but the logs stop there; it appears to timeout and then rollback). It mentions "service logs" but I do not see any logs in CloudWatch with any additional details.

```09-21-2022 02:23:38 PM [AppRunner] Service status is set to RUNNING.
09-21-2022 02:23:37 PM [AppRunner] Service update failed. For details, see service logs.
09-21-2022 02:17:00 PM [AppRunner] Performing health check on port '3000'.
09-21-2022 02:16:59 PM [AppRunner] Provisioning instances and deploying image.
09-21-2022 02:16:54 PM [AppRunner] Service status is set to OPERATION_IN_PROGRESS.
09-21-2022 02:16:53 PM [AppRunner] Service update started.
```

Note: I tried adding a dummy variable that has no effect on the app, and the same result occurred. Something appears to be broken here on App Runner.

Answer

The bug: if you change anything (or nothing) in the configuration using the console GUI, ***App Runner will silently change the HTTP endpoint to `/`***!

Perhaps the log statement thinks printing out `/` is redundant, but it is not because the log message appears identical to the original default *TCP* check.

I got around this issue by adding a redundant health check on `/`.

App Runner fails to update on any environment variable change

Relevanter Inhalt