[Kubernetes] Update OOMKill OOTB Dashboards

[triviajon]

What does this PR do?

Updates the "Containers OOM Killed" widget on the Kubernetes Pods and Kubernetes Clusters OOTB dashboards. Replaces the old graphed metric with clamp_min(diff(restarts), 0) * last_state.terminated{reason:oomkilled}, using restarts as the event trigger and last_state.terminated as an "OOM filter".

This also meant that we need to group by both pod_name and kube_container_name, because otherwise we get situations where a pod with many containers can see massive overcounting for a single OOM kill (n OOM kills can be overcounted to n(n-1)/2).

The fractional values sometimes visible are a known artifact of the kubelet check not reporting restarts during a container's restart window, which means that there is a small metric gap. Thus taking the diff/derivative over that spreads the increment in restart over multiple points.

Motivation

CONTINT-5192, FSCONS-1115

Review checklist (to be filled by reviewers)

• Feature or bugfix MUST have appropriate tests (unit, integration, e2e)
• Add the qa/skip-qa label if the PR doesn't need to be tested during QA.
• If you need to backport this PR to another branch, you can add the backport/<branch-name> label to the PR and it will automatically open a backport PR once this one is merged