feat(infra): show allocatable capacity and utilization for Kubernetes nodes - #1390

🟢 Confidence 9/10 · safe to merge
SQL, spec mapping and formatting are pinned by unit tests and the SQL baseline; only the collector's allocatable metric names are unverifiable here.
_{quality 100/100 · no findings · tests covered · risk medium}

Adds allocatable CPU/memory capacity and usage/allocatable utilization to the Kubernetes Nodes list and node detail: the fractions are computed in ClickHouse, mapped through the domain schemas, and rendered as meters with used/allocatable captions. Contained, consistent across the stack, and safe to merge.

• listNodesQuery and nodeDetailSummaryQuery now scan k8s_cluster allocatable metrics alongside kubeletstats usage
• nodeGaugeTimeseriesQuery divides a usage metric by a capacity metric per bucket
• Node list/detail show utilization meters, sort by CPU utilization, and gain CPU/memory chart pivots
• k8s.kubelet.version, k8s.node.uid and the cluster name are read with anyIf(attr != ''), so they no longer come back blank

_{613237d · Updated on every push. Reply "won't fix" to dismiss a finding, or mention @maple-review-bot to ask about one.}