starrocks-admin-cluster-health

What problem does it solve?

This skill solves the problem of diagnosing StarRocks cluster instability by helping you quickly assess FE/BE/CN liveness, tablet health, replica status, compaction backlog, and storage imbalance.

Core Features & Use Cases

• FE quorum and metadata health checks: identifies down frontends, validates quorum availability, and checks metadata edit-log replay lag to catch control-plane issues early.
• BE health, disk pressure, and capacity risk: reports dead backends and evaluates disk utilization thresholds to predict write stalls and emergencies.
• Tablet, replica, compaction, and balance triage: finds unhealthy tablets/replicas, surfaces compaction backlog pressure, and checks ongoing balance operations to confirm whether rebalancing is working.
• Prometheus alert readiness: provides metric names and alert rules to turn health signals into actionable monitoring.

Example use case: after a BE crash you need to confirm which replicas are missing, whether repair is in progress, and whether compaction and disk usage are likely to degrade query performance.

Quick Start

Use the starrocks-admin-cluster-health skill to run FE and BE health checks and return a prioritized diagnosis for FE quorum, disk usage, unhealthy tablets, replica problems, compaction pressure, and cluster balance status.