Curious how people handle this in practice.
Say I’m looking at a latency spike in my observability tool and grouping it by k8s.deployment.name. One deployment is clearly the outlier.
At that point I usually want to know what was happening underneath it: which pods were running, whether one restarted, whether there was a rollout, what nodes those pods were on, whether a ConfigMap changed, what Kubernetes events fired, etc.
The annoying part is that most of that information exists, but I end up leaving the chart and reconstructing the context manually with kubectl, the Kubernetes dashboard, events, and logs.
How are people doing this today? Do you have something that preserves the relationship between application telemetry and the Kubernetes objects behind it, or is jumping between tools basically the normal workflow?