Health Checking Issues and Confusion

[Marco Galassi]

Hi everyone!

Apparently we are experiences issue with the Health Checking system of App Engine.

We have disabled the regular health check mechanism as explained in this doc page. And, in fact, we do not receive any logs of requests pointing to /_ah/health.

But if we filter by nginx.health_check logs we see logs for health check requests to /_ah/vm_health.

First of all, we do not know what this is exactly.

Second, we see a long sequence of requests to /_ah/vm_health that are answered 200 and then suddenly it become 503. And after some of them, like 15ish requets

that answered 503, the server reboots.

What could be the causes of this? Where can I point my eye to search for answers and possibly solutions?

Any help is very appreciated.

Thank you very much

[Shivam(Google Cloud Support)]

After health checks are disabled, /_ah/vm_health path is created for backend health checks. As far as I know, it should return 200. Issues such as these most of the times tend to be project/application specific. I would recommend to post on App Engine public issue tracker so we can take a look at this at project level.