Kubernetes 24/7 AI SRE with K8sGPT Operator

If you've tried and love K8sGPT, now we're bringing you K8sGPT-operator, the in-cluster evolution of the project. With this, you're now able to have 24/7 monitoring with Observability signals generated to hook it up to your existing infrastructure.
Links:
k8sgpt.ai
github.com/k8s...
github.com/k8s...

[@Mandalorx10]

Well damn this looks amazing!

[@michaelbacchiocchi8111]

Looks really powerful in gaining meaningful insights into complex clusters. Your out-of-memory example was quite instant - is that a due to setting frequent polling intervals? How would that scale? Is the state data memory resident? If so, how does recoverability factor in?

[@rsjoat]

can it analyze historical changes to manifests (or api sequences for all native and CRs) since cluster bootstrap and predict how changes affected resource utilization ?
like: ingest and compare all known best practice designs and find ways to better optimize for cost and performance?
how do you plug into local bare metal ai ?