[dholt]

Summary

Refresh published component versions and keep offline preparation aligned with role defaults:

• Slurm 26.05.5; GPU Operator v26.7.2; device plugin and GPU Feature Discovery 0.20.1; MIG Manager 0.15.1.
• Prometheus 3.15.0; Alertmanager 0.34.1; Grafana 13.2.4; node-exporter 1.12.1.
• Distribution registry 3.1.2; registry proxy 0.7.0; Network Operator 26.7.0.
• Retain kube-prometheus-stack 88.6.5; major-chart migration is outside this change.
• Align offline chart, image and archive examples; add regression tests for pin consistency.
• Preserve the Operator-managed driver pin at 580.126.20 rather than changing driver branches with the chart update.

Validation

On head a1e1f4cbb7562d5dabf000baaf6445e379240a75:

• Independent code review approved.
• Role lint passed for 341 files; 12/12 playbook syntax checks and 134/134 local validation tests passed.
• Offline pin checks, MIG package mappings and both monitoring example renders passed.
• Single-node Kubernetes, GPU Operator and a GPU pod running nvidia-smi passed with an A40 and preinstalled driver 580.178.04. This does not validate the Operator-managed driver pin.
• Nonpersistent monitoring deployed successfully; Prometheus collected GPU, node and Kubernetes metrics. Grafana loaded 30 dashboards, and sampled panel queries returned data.

Remaining validation

This PR is intentionally draft, not release-ready:

• Slurm 26.05.5 build and GPU job smoke tests were not run.
• Persistent monitoring, upgrade paths, live MIG reconfiguration and Network Operator secondary-network traffic are not established by these results.
• Some existing GPU Nodes host/node panels returned no data; the GPU metrics queries were checked separately. This is not a claim that every panel is validated.
• Network Operator 26.7.0 requires Kubernetes 1.32 or newer; older cluster-version overrides need separate consideration.
• No new operating-system, firmware or hardware support is claimed.