What problem does it solve?
This Skill reduces operational toil by automatically detecting and remediating common StarRocks health issues in Routine Load ingestion, tablet distribution, compaction backlog, stale statistics, and partition/load-label cleanup.
Core Features & Use Cases
- Auto-Restart Routine Load (transient PAUSED): Scans Routine Load jobs and automatically resumes only those paused due to transient reasons (e.g., connectivity or broker availability), while flagging data-quality or Kafka-retention issues for manual intervention.
- Compaction backlog remediation: Triggers administrative compaction to address tablets with high compaction score when they exceed a configurable threshold.
- Data freshness and storage hygiene: Refreshes stale optimizer statistics via ANALYZE based on job age/status, drops expired partitions in a retention window (dry-run first), and reports cancelled load labels for safe retry workflows.
Quick Start
Ask the agent to run a full health check for your StarRocks cluster and provide a prioritized action report (auto-resumes, compaction triggers, analyze tasks, and partition retention dry-run results) for the databases you specify.
