Comment 42 for bug 2056193

One thing to mention on the hypothetical is that because httprequest-lego-provider/0 is the same unit after it was removed and redeployed - we only redeployed because we were seeing the error reported in this bug. In other words, we were seeing this problem before the redeployment as well. We had some errors in the application when we initially deployed it (per the initial bug report) and we tried to run actions while it was in that state and they were blocked, but then remained blocked after that error was fixed so we're not sure if that was related in any way.

There was no change from k8s to machines. I tried to confirm there were no k8s resources related to this application after the application removal (kubectl get all | grep httprequest-lego-provider) before redeploying.