benchmarking: add OOM kills to server telemetry

[Nishanth Kotla (Nishanth29)]

Depends on #2221 and #2386. Only the last commit, 24368cd8 ("benchmarking: add OOM kills to server telemetry"), is new; I'll rebase once those merge.

Part of #1590

Adds an oom_events block to server telemetry:

• node, pod: OOM kills over the run, from cAdvisor's container_oom_events_total. node counts global OOMs. pod counts kills inside a worker pod (actor sandboxes and the ateom container) and is best effort: a kill whose cgroup is gone before the scrape is missed.
• worker_restarts, worker_oomkilled_pods, evicted, worker_pods_lost: from the Kubernetes API, read before and after the run. evicted comes from Evicted events, as atecontroller deletes an evicted pod at once.

Other changes:

• monitoring.yaml keeps container_oom_events_total.
• The runner's benchmark-workloads Role can also list events.
• The README explains what each field does and doesn't catch.

Live check (gVisor, 3m):

• Actor limit 256Mi, 2 users: pod = 6, matching 6 kernel oom-kill lines, one per actor sandbox. Other fields 0; the worker stayed up.

• No actor limit, 8 users: the kernel killed the whole worker twice at the pod cgroup limit (oom_memcg=/kubepods.slice), and the kubelet evicted it. evicted = 7 (1 eviction plus 6 replacement pods the kubelet refused to start), worker_pods_lost = 1, node = pod = 0, as documented.

• 80 measurements in the jsonl row.

• Tests pass

• Appropriate changes to documentation are included in the PR

[Nishanth Kotla (Nishanth29)]

cc Haowei Cai (Roy) (@roycaihw), Max Smythe (@maxsmythe) - ptal:)

[Nishanth Kotla (Nishanth29)]

Da Huang (@git286) Tim Bai (@baizhenyu), adding you as reviewers.

What this uses:

• Node-level OOMs and worker pod restarts/evictions: cAdvisor and the k8s API (outside Substrate).
• Kills inside the worker pod (pod, mostly actor sandboxes): cAdvisor too, since there's no Substrate OOM metric to read yet.
• Not caught: OOMs handled inside the sandbox (a microVM guest, or a single container hitting its own limit under gVisor).

If you'd rather add an actor OOM-kill counter in Substrate (e.g. from the oom_kill read in #2258), pod can switch to that or I can drop pod from this PR until then.