Azure/container-apps-deploy-action

Support waiting until deployment is completed

Opened this issue ยท 8 comments

Hi!

I am looking for a feature to be able to make this action wait until the deployment is actually completed. Now the action seems to be considered "completed" when the signal is sent away to Azure to initiate the deployment. I would like the action to wait until the new revision is deployed and is up and running OR until the deployment actually failed. It would be great to have this feature so that we don't have to implement it ourselves.

BR,
Rasmus

@razum90 Hey Rasmus, apologies for the delayed response -- this seems like something that we should be able to support! You're correct that we are firing off the az containerapp * command to either create or update the Container App, but we don't do anything additional after that with regards to the result of this command (pending an outright failure when calling the command). It makes sense to give users the ability to wait for the Container App deployment to be completed -- I'll take a look at adding a flag to signal the action to wait for a completed deployment and what az containerapp calls we'll need to make for the polling flow. Thanks for the request!

Sounds great @cormacpayne. Let me know if you want any help to test it.

Hey all, we have been looking for this exact feature as well. @cormacpayne, is there an update on when this flag will be released as a part of the action?

Any updates? This is a critically important feature for us.

Phiph commented

Yeah I'd like to see this too!

It seems to me that this should be prioritized.

Today I tried this for the first time, was very happy to see it pass, and then spent a while troubleshooting why I got a weird error message that should not happen. And it turned out the deploy had eventually failed as the instance didn't start. This seems like a crucial feature for anyone unless you want to always cross your fingers and hope everything is alright without know so.

Thanks everyone for your comments. I confirm that this change is now part of our backlog (+ @snehapar9 for FYI).
It actually seems to me that we could make this the default behavior, as I agree it makes sense to wait for the container app to be deployed before exiting the task.

I support the request.

Currently the job is considered green, even if the revision fails. It would be nice to not use my hacky custom workarounds :D