메뉴
Processes and analyzes TiDB profiler zip packages (CPU or heap) for components like TiDB, TiKV, PD, or TiFlash. It unzips, aggregates the profiler data, and uses `go tool pprof` to report on the most time-consuming functions (CPU) or highest memory allocations (heap).
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | tidb-profiler-analyzer |
| description | Processes and analyzes TiDB profiler zip packages (CPU or heap) for components like TiDB, TiKV, PD, or TiFlash. It unzips, aggregates the profiler data, and uses `go tool pprof` to report on the most time-consuming functions (CPU) or highest memory allocations (heap). |
Your primary goal is to organize and analyze profiler data from various TiDB component (tidb, pd, tikv, tiflash) .zip archives.
Follow these steps meticulously:
.zip files present in the current working directory. Each zip file should be unzipped into a separate, appropriately named directory.'profiler_data' at the root of the current working directory. This directory will house all collected profiler files.'profiler_data' directory.'profiler_data' directory (CPU or Heap). You can often infer this from filenames (e.g., containing "cpu" vs "heap"). If unsure, you may need to ask the user.
b. Use the go tool pprof to analyze the collected data.
* For CPU profiles: Run go tool pprof -top <path_to_profiler_file> to find the most time-consuming functions.
* For Heap profiles: Run go tool pprof -top --alloc_space <path_to_profiler_file> to find the functions with the most memory allocations.
c. Present a summary of the analysis results to the user..zip files and the now empty unzipped directories to tidy up the workspace.Fix software bugs with root-cause analysis, minimal diffs, regression tests, and a mandatory self-check loop. Use for general bugfix work that is not covered by a more specific domain skill.
Fix TiDB optimizer bugs with minimal diffs, hypothesis-driven validation, and regression tests. Use the bundled issue precedents only to avoid re-fixing already solved bugs or inspect suspect historical fixes.
Gather PR review context: all comments (including review suggestions), diff context, and CI status; download GitHub Actions logs on failures. Prefer GitHub MCP for reads, fall back to gh CLI when needed.
Use for projects, tasks, or code that involve context engineering and context management for AI agents: keep prompt prefixes stable for KV-cache, use append-only context, prefer tool masking over tool removal, offload large observations into filesystem memory, recite goals/todos to control attention, preserve errors for recovery, and avoid few-shot pattern lock-in. Use when building or debugging agent loops, prompt/context schemas, memory strategies, or tool-availability policies.
TiDB-specific documentation lookup guided by this repo's llms.txt hub: read llms.txt, pick the best linked source (tidb-dev-guide llms-full.txt, TiDB user guide llms.txt, or Uber Go style guide), fetch it, then search within it to answer with precise references. Use when asked to find docs, research, or provide doc-grounded answers about TiDB development or usage.
Expertise in adding new test cases for the TiDB plan replayer. Use when the user provides a plan replayer zip file and wants to create a new test.