Motivation
The block cache consumes memory, so for clusters where it is enabled we need to watch available/headroom memory to catch memory pressure introduced by the cache and to inform the default sizing policy.

Approach

• For clusters reporting the block cache as enabled (via the enabled stat), build a Grafana panel showing available system memory over time.
• Correlate with the block cache size/count stats already in the blockCache ServerStatus section.

Definition of Done

• A Grafana panel exists showing available memory on clusters with the block cache enabled.
• The panel is scoped to enabled clusters and refreshes on the standard fleet cadence.

Note: this is an observability/cloud-monitoring task and may need a cloud/observability owner; it is tracked here so the full production effort lives under one epic.