kali20gakki/msAgent / skills/host-cpu/msagent-perf-cpu-config-audit
msagent-perf-cpu-config-audit
Verified昇腾训练/推理主机的环境配置体检:对 NUMA 亲和、大页/THP、governor 频率、cgroup 限流、IRQ smp_affinity、环境变量等 8 个配置域做单点快照审计,输出分级 Finding 与可回滚修改建议,不改动运行状态。触发:检查环境、配置体检、新机器开箱、换容器、 大页没配、governor 不对、cgroup 限流、中断亲和配置、NUMA 配置检查、绑核没生效。 不用于:训练变慢/step 抖动/利用率低的根因归因(→ perf-cpu-perf-triage); 产出绑核方案(→ perf-cpu-bind-tune);运行期实际落核观测(→ perf-cpu-tracing)。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 5.2 KB |
| references/check-items.md | markdown | 6.5 KB |
| references/process-roles.md | markdown | 1.7 KB |
| references/profiles/ascend910.json | config | 1.6 KB |
| references/profiles/kunpeng920-910b.json | config | 507 B |
| references/rules-locality.md | markdown | 2.3 KB |
| resources/report_template.html | other | 11.5 KB |
| scripts/audit.py | script | 19.6 KB |
| scripts/baseline.py | script | 4.4 KB |
| scripts/checker.py | script | 39.6 KB |
| scripts/classify.py | script | 2.7 KB |
| scripts/discover.py | script | 6.6 KB |
| scripts/npu_query.py | script | 15.0 KB |
| scripts/preflight.py | script | 3.3 KB |
| scripts/procfs.py | script | 9.0 KB |
| scripts/report.py | script | 6.4 KB |
| scripts/rules_locality.py | script | 8.7 KB |
| scripts/static_collect.py | script | 13.8 KB |
| scripts/topology.py | script | 9.4 KB |
| scripts/workspace.py | script | 2.1 KB |
scripts/report.py
View on GitHub# -------------------------------------------------------------------------
# This file is part of the MindStudio project.
# Copyright (c) 2026 Huawei Technologies Co.,Ltd.
#
# MindStudio is licensed under Mulan PSL v2.
# You can use this software according to the terms and conditions of the Mulan PSL v2.
# You may obtain a copy of Mulan PSL v2 at:
#
# http://license.coscl.org.cn/MulanPSL2
#
# THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND,
# EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT,
# MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE.
# See the Mulan PSL v2 for more details.
# -------------------------------------------------------------------------
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""report.py — HTML 分节报告 + 终端摘要 + findings.json 落盘。
HTML 由 resources/report_template.html 模板 + JSON payload 渲染,
不拼接 HTML 片段(模板占位符 __PAYLOAD__ 注入)。
"""
from __future__ import annotations
import json
import os
from typing import Any, Dict, List
# 域中文名
DOMAIN_NAMES = {
"numa": "A · NUMA 与拓扑", "bind": "B · 绑核与 CPU 隔离",
"hugepage": "C · 大页与内存", "freq": "D · 频率与电源",
"irq": "E · 中断与软中断", "cgroup": "F · cgroup 与配额",
"env": "G · 运行时环境变量", "npu": "H · NPU 适配",
}
_STATUS_ORDER = {"CRIT": 0, "WARN": 1, "FACT": 2, "OK": 3}
def _short(value: Any, limit: int = 40) -> str:
"""终端摘要取值:dict/列表截断展示,避免 HTML 转义(终端非 HTML 上下文)。"""
text = value if isinstance(value, str) else json.dumps(value, ensure_ascii=False)
if len(text) > limit:
return text[:limit] + "…"
return text
def build_payload(results: List[Dict[str, Any]], findings: List[Dict[str, Any]],
suspects: List[Dict[str, Any]], meta: Dict[str, Any],
baseline_diff: List[Dict[str, Any]],
actions: List[Dict[str, Any]] = None) -> Dict[str, Any]:
"""构造注入模板的 payload。"""
# 按域分组 + 排序(CRIT → WARN → FACT → OK)
# OK/FACT 只进折叠表(项/检查项/判定/当前值),去掉仅供解释的字段,控制报告体积
domains: Dict[str, List[Dict[str, Any]]] = {}
for result in results:
if result["status"] in ("OK", "FACT"):
result = {k: v for k, v in result.items() if k not in ("explain", "advice")}
domains.setdefault(result["domain"], []).append(result)
for domain in domains.values():
domain.sort(key=lambda r: (_STATUS_ORDER.get(r["status"], 9), r.get("id", "")))
counts = _status_counts(results)
top5 = _top5(results)
return {
"meta": meta, "counts": counts, "top5": top5,
"domains": {DOMAIN_NAMES.get(k, k): v for k, v in domains.items()},
"domain_keys": [DOMAIN_NAMES.get(k, k) for k in domains],
"findings": findings, "suspects": suspects,
"baseline_diff": baseline_diff,
"actions": actions or [],
}
def _status_counts(results: List[Dict[str, Any]]) -> Dict[str, int]:
counts = {"CRIT": 0, "WARN": 0, "FACT": 0, "OK": 0}
for result in results:
if result["status"] in counts:
counts[result["status"]] += 1
return counts
def _top5(results: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
bad = [r for r in results if r["status"] in ("CRIT", "WARN")]
bad.sort(key=lambda r: (_STATUS_ORDER[r["status"]],
-{"high": 3, "medium": 2, "low": 1, "": 0}.get(r.get("benefit", ""), 0)))
return bad[:5]
def render_html(payload: Dict[str, Any], template_path: str, output_path: str,
single_file: bool = False) -> str:
"""模板 + payload → hostcpu_config_report_<host>_<ts>.html。"""
with open(template_path, "r", encoding="utf-8") as fh:
template = fh.read()
safe_payload = json.dumps(payload, ensure_ascii=False).replace("</", "<\\/")
rendered = template.replace("__PAYLOAD__", safe_payload)
if single_file:
rendered = rendered.replace("__PAYLOAD__", safe_payload)
parent = os.path.dirname(output_path)
if parent:
os.makedirs(parent, exist_ok=True)
with open(output_path, "w", encoding="utf-8") as fh:
fh.write(rendered)
return output_path
def save_findings(findings: List[Dict[str, Any]], actions: List[Dict[str, Any]],
path: str) -> str:
data = {"findings": findings, "actions": actions}
parent = os.path.dirname(path)
if parent:
os.makedirs(parent, exist_ok=True)
with open(path, "w", encoding="utf-8") as fh:
json.dump(data, fh, ensure_ascii=False, indent=2)
return path
def terminal_summary(results: List[Dict[str, Any]], findings: List[Dict[str, Any]],
meta: Dict[str, Any]) -> str:
"""终端摘要(对话内直接输出):得分 + Top 问题 + 复检建议。"""
counts = _status_counts(results)
lines: List[str] = []
score = _score(results)
lines.append(f"主机配置审计摘要 · {meta.get('host', '?')} · 模式={meta.get('mode', '?')}")
lines.append(f"环境得分:{score}/100 "
f"CRIT {counts['CRIT']} WARN {counts['WARN']} "
f"OK {counts['OK']} FACT {counts['FACT']}")
for item in _top5(results):
lines.append(f" [{item['status']}] {item['id']} {item['name']}:"
f"{_short(item.get('value_text') or item['value'])}"
f"(阈值 {_short(item['threshold'])})")
for finding in findings:
lines.append(f" [规则 {finding.get('rule')}] {finding['problem']}"
f"(suggested_cpuset={finding.get('suggested_cpuset') or '-'})")
lines.append(f" OK: {counts['OK']} 项通过")
lines.append("复检建议:修复后同参数复跑本审计对比 baseline。")
return "\n".join(lines)
def _score(results: List[Dict[str, Any]]) -> int:
"""环境得分(CRIT/WARN 加权,仅作趋势指标,不替代逐项判定)。"""
counts = _status_counts(results)
penalty = counts["CRIT"] * 20 + counts["WARN"] * 5
return max(0, 100 - penalty)
def suspect_items_json(suspects: List[Dict[str, Any]], path: str) -> str:
parent = os.path.dirname(path)
if parent:
os.makedirs(parent, exist_ok=True)
with open(path, "w", encoding="utf-8") as fh:
json.dump(suspects, fh, ensure_ascii=False, indent=2)
return path
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |