We’re mostly on Azure with about 30 subscriptions and two AKS clusters and this year cost ended up on my plate. I spent most of last quarter working through the advisor cost tab, resizing the VMs it flagged, deleting orphaned disks and buying the savings plan it kept recommending and after all of that, the bill only moved about 4%.
That’s gotten me to finally open the usage blade on the Log Analytics workspace both clusters send to. ContainerLogV2 is the biggest table by a long way with Perf and KubeEvents behind it and the workspace now costs more than the node pools it’s monitoring. As far as I can tell, Container Insights is still on whatever settings were picked when the clusters were created.
I can fix the logging but I only found it by poking around myself and I’m guessing it isn’t the only thing Advisor misses. What else have you found in AKS that Advisor never flagged?