Daniel Clark

This is now fixed and available in GKE

Change captured in https://github.com/GoogleCloudPlatform/prometheus-engine/commit/30b003b8b91bc4a8654c3527ed5d996024711b24

Hi @superbrothers thanks for mentioning (and sharing) arm64 images! We actually have support for building them yourself using Google Cloud Build from #346. You should be able to build any...

Whoop - we have support for ARM images as of https://github.com/GoogleCloudPlatform/prometheus-engine/releases/tag/v0.7.0. Closing.

@willchan - when you mention the "default VerticalPodAutoscaler" in autopilot clusters, where is that documented? When I create an autopilot cluster, i don't see any provisioned ```bash % k get...

Yes, that's my understanding as well. Deploying GMP collectors as a DaemonSet mitigates what we've been calling "shard-overload", but it doesn't fix the problem entirely. It's completely possible to overwhelm...

> That said, from VPA's perspective, perhaps this is a bug in the collector. Why does the memory usage vary by Pod/Node? Because a Prometheus scraping 10k time series uses...

Hi @alpenfritz - apologies for the delayed response. Indeed this query function is not yet supported. We are aware of this and will post here with any updates in support....

Update here: we have 1.25+ support on Autopilot clusters. Will keep this issue open as we work on 1.24...

Quick demo: https://github.com/k8sgpt-ai/k8sgpt/assets/3731548/ee72cdcd-0fdd-42fd-be2c-fdd1a397fb9f The output isn't perfect. I'm betting we can improve with a better prompt...