Stopping a function runtime container manually using CLI and triggering again is not possible

Hi there,

so I am using several appwrite functions and because there is a delay on the first run, I was thinking, I would like to test a certain use case when the function did not have the initial first trigger. So the use case when the docker container of the function is not running. So I thought I could simply stop the container:

docker container stop <function-container-name>

and then retrigger the function. But in this case I get an error and only a complete restart (docker compose down and docker compose up -d) of the appwrite docker compose fixes this issue. I thought: Yeah maybe this is something I shouldn't do because the appwrite-executor is managing the container and I shouldn't tamper with the function container directly. But what happens if a container fails? Is a restart of the appwrite function impossible and a complete restart either of the appwrite docker compose (or maybe the appwrite-executor) is necessary? This would lead to a downtime.

So my actual question is: Is this behaviour by design?

The runtime container shouldn't just fail ..

You'll need to restart the executor yes

I agree it shouldn't fail but isn't it a possible case? I was wondering which behavior is set for the restart property of the docker container of the functions. In case it is set to on-failure I guess it might be sufficient.

But in general, is it possible to stop a function container in a clean way? I have set my function threshold to never stop for performance reasons.

"RestartPolicy": {
  "Name": "no",
  "MaximumRetryCount": 0
}

So it is set to no. Is there a reason which I am missing? Shouldn't on-failure and unless-stopped be the better configuration, especially if the restart of appwrite-executor is necessary for a on-failure case?

Maybe but it doesn't happen in my experience

Best thing to do to stop a container would be to wait for it to be killed to die inactivity.