Mark Summers

I think I've found the cause of the issues: The container CPU limit is being returned as 2 (presumably due to default Linux cgroup behaviour) when it's actually 0 in...

After experimenting with different CPU limits on individual containers, I've realised that the values returned for ecs_task_percpu_usage_ratio are inexplicable. I would expect them to be affected by the same issues...

I did something similar using `redir` and running both services under `supervisord` (package name: supervisor). It seemed to work quite well. Probably, `socat` will also work with the right options...

Possibly related to https://github.com/runatlantis/atlantis/pull/4192#issuecomment-3204874755. If I understand correctly, the locks are not really so isolated as the first read of the code would suggest.

@tzneal I have a nodeSelector based on a custom label, which has been applied to the ebs-csi-node DaemonSet. The idea is that the CSI node pod should only run on...