Diagnose TiKV memory growth, OOM crashes, and memory hotspots.
npx skills add https://github.com/alex-quan-001/nutshell-skills --skill tikv-memory
Please help me install this Agent Skill. Skill: tikv-memory Source: https://github.com/alex-quan-001/nutshell-skills/tree/main/skills/diagnosis/tikv-memory Command: npx skills add https://github.com/alex-quan-001/nutshell-skills --skill tikv-memory
TiKV memory issues are notoriously difficult to troubleshoot due to its multi-module architecture and lack of unified memory accounting, leading to wasted time guessing at root causes of gradual memory growth, sudden OOM crashes, and memory hotspots.
Use the tikv-memory skill to investigate the gradual memory growth and OOM crashes affecting your 3-node TiKV v7.5.6 cluster, and retrieve the recommended workaround for the pessimistic lock memory leak pattern.
High-intent search queries and answers about installing and using this skill.
TiKV OOM crashes often stem from resolved-ts memory pressure, coprocessor backlogs, or pessimistic lock leaks. You can root cause TiKV memory growth by routing initial checks, reviewing key configurations, and analyzing per-module metrics to pinpoint the exact memory consumer.
To collect a TiKV heap profile for diagnosing memory hotspots, follow the specific profile collection procedures designed for pre-v7.5 and v7.5+ TiKV instances, which help identify the exact modules consuming excessive memory.
Common known TiKV memory bloat issues include scheduler lock_key leaks, pessimistic lock memory accumulation, resolved-ts OOM, and coprocessor task backlogs. A known issue reference library provides documented workarounds and fixed versions for these verified bugs.
To check if TiKV memory pressure is caused by excessive region count, review key metrics for per-module memory estimation and analyze TiDB Dashboard metrics to identify if region overhead is the primary memory consumer triggering swap anomalies or OOMs.
Yes, this diagnostic approach applies to TiKV clusters experiencing repeated OOMs after a scale-in operation by cross-referencing symptoms against known scaling-related memory issues, checking region count metrics, and applying the correct mitigation steps.
mohitmishra786Profile heap memory allocations on Linux with heaptrack.
openeuler-mirrorCorrelate kernel logs, process memory snapshots, and cgroup metrics to diagnose Linux OOM events.
alex-quan-001Evaluate HPA autoscaling quality for TiDB Cloud components and generate HTML reports.
4lermanProfile Node.js, Python, and Go applications for CPU, memory, and I/O bottlenecks.
