fix：DDC meta service health aggregation

[Al-assad]

Motivation

kubectl get ddc can still show cluster health as green after one MetaService pod is down. The DDC cluster health aggregation currently only considers FE availability and compute group availability, so MetaService degradation is not reflected in status.clusterHealth.health.

There is also a stale status path: once MetaService Phase becomes Ready, UpdateComponentStatus does not downgrade it when the available MetaService replicas drop below the desired replicas.

Changes

• Downgrade DDC MetaService Phase to Reconciling when not all desired MetaService replicas are ready on the latest StatefulSet revision.
• Include DDC MetaService availability and phase in cluster health aggregation:
  • MetaService unavailable -> red
  • MetaService available but not fully ready -> yellow
• Add unit tests for MetaService phase downgrade and DDC cluster health aggregation.

Tests

go test ./pkg/controller/sub_controller/disaggregated_cluster/metaservice
go test ./pkg/controller -run TestReorganizeStatusConsidersMetaServiceHealth
go test ./pkg/controller/sub_controller/disaggregated_cluster/...