resource-isolation

StarRocks queries fall into default_wg when resource group classifiers are missing or mismatched. Diagnose this by analyzing audit logs to identify unclassified queries and adding the correct classifiers to enforce proper workload isolation.

To stop long-running analytical queries from bypassing limits and triggering spilling, inspect BE metrics and tune big query circuit breakers. Adjusting resource group limits and queue sizing prevents memory spill contention and restores stable latency.

The best way to handle StarRocks query queue full rejections during traffic bursts is to review and adjust query queue sizing. Configuring resource group limits and big query circuit breakers safely absorbs bursts without starving other workloads.

Prevent statistics collection jobs from overwhelming StarRocks BRPC worker threads by diagnosing statistics collection saturation. Inspect FE and BE metrics to identify the saturation point and apply emergency configuration changes to restore fair scheduling.

Yes, you can use a SQL blacklist to block specific big queries from starving StarRocks workloads. Governance tuning guides the configuration of SQL blacklists alongside resource group limits and circuit breakers to enforce query isolation.

Validate StarRocks workload isolation and latency recovery by confirming query classification and limit enforcement. Analyze audit logs and FE metrics to verify that queries are no longer falling into default_wg and fair scheduling is restored.