Which GitLab endpoint should be used as a health check
Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.
Hello,
On v12.9.3 and lower, we were using the /explore page to monitor for container health for the GitLab application.
In our upgrade to v12.10.3, we found that /explore now redirects to /users/sign_in, which returns a 302 status and not a 200 status, thus failing the health check. It was advised to try using /users/sign_in, however that was returning a 302:
[08/May/2020:20:36:47 +0000] "GET /users/sign_in HTTP/1.1" 302 151 "" "ELB-HealthChecker/1.0"
This is related to this change: !22522 (merged)
I tried using /users/sign_in on our production and a brand new stage environment with the same result. I ended up using /robots.txt for health, however, @twk3 mentioned (omnibus-gitlab#5326 (closed)) that's not a good use case to test the rails application.
Is there a better page to use for an unauthenticated user to test container health? It would need to return a 200 status code.
Thanks, Tom