obdiag-performance

Diagnose OceanBase host load and slow queries by aggregating ASH, AWR, perf flame graphs, sysstat, and disk IO data to pinpoint system bottlenecks. It correlates server-side performance metrics with host-side load to identify root causes of latency and session buildup.

Identify CPU and system call hot spots in OceanBase by collecting and analyzing perf flame graphs. This visualizes CPU consumption across active threads to reveal which internal functions or system calls consume the most processing time during peak load.

Collect ASH and AWR data for active session analysis and historical workload insights in OceanBase. Running gather_ash captures current active session activity, while gather_awr generates historical workload reports to establish performance baselines.

OceanBase performance diagnostics works with disk IO monitoring by running tool_io_performance to characterize IO wait and throughput. It aggregates sysstat host baseline metrics to correlate disk IO bottlenecks with overall host load and query latency.

Run root cause analysis alongside host performance diagnostics when OceanBase latency, session buildup, or high host load persists after initial bottleneck identification. Coordinating rca_run with check_cluster validates cluster configuration and isolates systemic deployment issues.