Comment 35 for bug 2056193

Ok, I tried redeploying the application (I did a `juju remove-application httprequest-lego-provider`, waited til that had completed, and confirmed there were no k8s resources associated with that application in the k8s namespace for the model). Unfortunately actions are still stuck. If I run `juju operations` I see previous operations I've tried to execute on the model, including those from the previous iteration of the httprequest-lego-provider charm. I assume that's expected, but just mentioning it in case.

I also tried running an action while running `juju debug-code` on the application, but that didn't seem to work (it didn't intercept the charm code at all). I also tried adding some logging to the charm to notify when various parts of the code are triggered, but from doing so it seems like when running the action the charm's __init__ method isn't even called (which possibly explains why `juju debug-code` doesn't intercept it).

I'm kind of out of ideas here, apart from potentially trying to deploy it under a different name. I'd really appreciate a live debugging session with someone who understands the internals of this.