Moving it to the rook.

I thought that 4.9 still had the uninstall/reinstall bug where the rook operator pod needs to be restarted to fix this. Try restarting the rook operator pod to see if the issue resolves itself.

The customer could also disable the health checks for the ObjectStore. That will also remove the readiness probe from the RGW, but if there is only one, that shouldn't be an issue. Backporting this to 4.9 will prove to be a bit of a challenge, and too many competing priorities keep coming up.

Forgot needinfo last week

After further discussion, backporting it will be difficult with other related rgw readiness probe issues. 
Blaine, to what release have we already backported the bucket health check disabling? If that's sufficient, please close it.

Right. I believe ODF 4.10 has a number of additional patches, and I believe this issue should not be possible there. In 4.13, the health check routine is removed entirely after hearing concerns from the RGW team.