Falcon Taylor-Carter

I am using ephemeral runners, as well as the GKE cluster-autoscaler. I'm also using RunnerDeployments. I'll try out `RUNNER_FEATURE_FLAG_EPHEMERAL` and see if it improves things

We're still occasionally seeing this issue. I'll post more info on it when I have it but I haven't noticed anything useful about it recently.

It could be yeah. I've also been seeing offline runners building up over time still. > actions-runner-controller garbage-collects any runner pods whose twins in GitHub API are offline for a...

Any updates here? Would really like to see this feature get through

This is a pretty important feature for this project to have support for 🙏

We're using only `batchRead` currently. The separate application instances are running on the same machine, which is part of what leads me to suspect it's something wrong with the client...

This is being deployed on GKE. We found that we reached the ~2K read limit on a 1 core pod, which reached high CPU, and we saw the same thing...

I'm also running into this issue, with EKS 1.21 and controller 2.4.4. If @iusergii got it to work, I assume it's fixed in EKS 1.22+ edit: Never mind, got it...

Any new plans for a fix here? This is a frustrating bug. At the very least it would be nice if we could remove the ``` Warning: Redundant empty provider...

+1 still seeing this