Add expireAfter field to cycle pods after a configured lifetime

[xiangyuwang-stripe]

Summary

• Adds expireAfter (a metav1.Duration) to HumioNodeSpec, configurable per node pool, that defines the maximum time a pod is allowed to run before being replaced
• Implements ensureExpiredPodsAreDeleted in the reconcile loop, which checks pod ages against the TTL and deletes expired pods oldest-first while respecting all existing safety gates (MaxUnavailable, MinReadySeconds, zone awareness, waitingOnPods)
• Runs as a separate reconcile step after spec-drift detection and before pod creation, so image/config updates always take priority over TTL cycling

Example usage

spec:
  expireAfter: "168h"  # cycle pods every 7 days
  # or per node pool:
  nodePools:
    - name: ingest
      spec:
        expireAfter: "24h"