Editing the ConfigMap / changing env vars does not force containers to restart

[batpad]

Currently, if you change an environment variable for one of the containers, of edit the ConfigMap, it does not automatically restart the affected containers - i.e. the environment variable does not propagate to the container with a deploy, and it needs a manual restart of the container for the environment variables to propagate.

Steps to Reproduce:

• Change an environment variable in values.yaml
• Deploy

Expected:

• My changed environment variable to be on the container

Actual:

• The container does not restart because it does not know the ConfigMap has changed.

From chat with @ranchodeluxe on Slack:

"personally I'd just make my CI/CD bounce all services at deploy time and then you don't have to worry"

That sounds good to me. Another option would be to use an annotation on the deployments that's a hash of the ConfigMap, so if the ConfigMap changes, the hash changes, and that would force a container restart.

This is not urgent, just wanted to drop the issue so as not to forget.

cc @ranchodeluxe @sunu

[emmanuelmathot]

To avoid a forceful restart, the solution involves creating a checksum based on the config map and including it in the deployment's metadata specification. This ensures that the pod is updated only when the config changes.