What problem does it solve?
This Skill eliminates the time-consuming process of manually searching scattered documentation and past oncall tickets to troubleshoot complex TiKV storage-engine, RocksDB, Titan, and disk-related incidents that cause cluster performance degradation, service unavailability, or capacity exhaustion.
Core Features & Use Cases
- RocksDB Incident Reference: Covers common RocksDB issues including compaction pressure, SST file corruption, pending compaction bytes overflow, manifest rewrite overhead, and WAL encryption gaps, with proven workarounds and fixed version information.
- Titan-Specific Diagnosis: Provides guidance for Titan GC-induced read/write slowdowns, inaccurate region size estimation, and severe data skew across TiKV instances caused by Titan enablement.
- Host and Disk Behavior Analysis: Includes reference material for host-level disk IO anomalies, RaftEngine periodic traffic spikes, and THP-related latency/memory regressions that impact TiKV performance.
- Use Case: When a TiKV cluster experiences sudden QPS drops, disk capacity exhaustion, or elevated write latency, use this Skill to quickly match observed symptoms to known root causes and proven mitigation steps.
Quick Start
Invoke the tikv-storage skill to diagnose the root cause of your TiKV cluster's abnormal disk usage growth and pending compaction bytes overflow, and retrieve the corresponding mitigation steps.