Include priorityClass in armada:queue:resource:(queued|allocated) rules

[JamesMurkin]

The underlying metrics have priorityClass in them, as we don't include it in the grouping it actually undercounts the values

This fixes the rules producing misleading data when queues use multiple priority classes

[greptile-apps]

Greptile Summary

This PR preserves the existing priorityClass dimension when aggregating queued and allocated resource metrics, preventing resources from different priority classes from being collapsed by max.

• Adds priorityClass to both aggregation stages of armada:queue:resource:queued.
• Adds priorityClass to both aggregation stages of armada:queue:resource:allocated.
• Existing in-repository consumers aggregate across the new dimension where appropriate.

Confidence Score: 5/5

The PR appears safe to merge with no identified correctness or security issues.

The underlying metrics already emit a bounded priority-class label, both aggregation levels now preserve it consistently, and the only in-repository downstream consumer explicitly sums across that dimension.

Important Files Changed

Filename	Overview
deployment/armada/templates/prometheusrule.yaml	Correctly retains the bounded priorityClass label in queued and allocated recording rules; no actionable regression was identified.

_{Reviews (1): Last reviewed commit: "Include priorityClass in armada:queue:re..." | Re-trigger Greptile}