---
name: component-failure-or-bug
description: Diagnose panic, restart, crash, known bug, and component failure root causes.
---

# Use When

Use this family subskill when the primary mechanism is component panic, restart, crash,
known bug, process failure, or version-specific defect.

# Required Evidence

- Panic stack, fatal/error signature, restart count, OOM kill, component count gap, known issue
  match, fixed release, or repeated version-specific bug signature.
- Evidence separating process crash from recoverable goroutine panic or benign alert.

# Classification

- `standard_root_cause_class`: `component_failure_or_bug`
- compatible `primary_class`: `component_restart_or_panic`
- common symptom tags: `panic_or_restart`, `oom_signal`, `slow_query`
- specialist skills: `tikv-panic`, `github-issue-search`, component skill

# Decision Rules

- Distinguish crash/restart, recoverable panic, known bug, and false alarm.
- For recoverable timeout panics, do not make the panic itself primary if upstream SQL/IO/capacity
  caused the timeout and no process crashed.
- Mention known issue only when evidence is strong; do not let issue search override local
  incident causality.

# Final Response Anchors

- For benign recoverable timeout panics, say `expected/benign` and `no TiDB fix needed for the
  panic itself` when supported.
- Name the stack/function, component, affected instance, and whether a restart occurred.