浏览 Trace 日志

SkillMonitoring & ops

Use when a Pipeline run fails or produces abnormal results and you need to read the Job's Trace logs, analyze the cause of errors, and provide fix suggestions. Trigger words: browse traces, job failure, troubleshooting run failures, trace logs, pipeline errors, job error.

Use 浏览 Trace 日志 in Claude, ChatGPT or Ahel Desktop

Free. Sign in, add 浏览 Trace 日志 and connect your AI. About a minute.

Also: Claude Code · Cursor · Codex

Then ask your AI: use the 浏览 Trace 日志 skill

Details

Instructions available. Your AI can read the instructions. Execution depends on the setup they require.

Add Ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

浏览 Trace 日志Start free

What this skill tells your AI

The instructions your AI receives, as published by forge-town/ordine in skills/ordine-browse-traces/SKILL.md and read by Ahel’s review.

概述

当 Pipeline 运行后 Job 状态变为 failed 或结果不符合预期时,需要通过 Trace 日志定位根因并给出修复方案。

诊断流程

第一步:获取 Job 状态

# 查看 Job 详情
curl -s http://localhost:9433/api/jobs/<JOB_ID> | python3 -m json.tool

关注字段:

字段说明
statusfailed = 运行失败,completed = 已完成(可能有部分错误)
error顶层错误信息(如果有)
result运行结果摘要
startedAt / completedAt计算运行耗时,判断是否超时

第二步:读取 Trace 日志

# 获取该 Job 的所有 Trace
curl -s http://localhost:9433/api/jobs/<JOB_ID>/traces | python3 -m json.tool

Trace 数据结构:

字段类型说明
idnumber自增 ID
jobIdstring所属 Job ID
level"info" | "warn" | "error" | "debug"日志级别
messagestring日志内容
createdAttimestamp时间戳

第三步:按级别过滤分析

优先查看 error 级别的 Trace:

# 只看错误日志
curl -s http://localhost:9433/api/jobs/<JOB_ID>/traces | \
  python3 -c "import sys,json; [print(t['message']) for t in json.load(sys.stdin) if t['level']=='error']"

再查看 warn 级别了解潜在问题:

# 只看警告日志
curl -s http://localhost:9433/api/jobs/<JOB_ID>/traces | \
  python3 -c "import sys,json; [print(t['message']) for t in json.load(sys.stdin) if t['level']=='warn']"

第四步:常见失败模式

模式Trace 特征修复方向
Operation 执行失败error 消息包含 Operation ID检查该 Operation 的 executor 配置(script/skill/prompt)
输入路径不存在error 消息包含 ENOENT 或 not found检查运行时传入的 -i 路径是否正确
Skill 调用超时长时间无新 Trace,最终 failed检查 Skill 服务是否在线,网络是否通畅
脚本权限不足error 消息包含 EACCES 或 permission denied检查脚本文件的执行权限
节点连接断裂warn 消息提示某节点未收到输入检查 Pipeline DAG 中节点之间的连线

第五步:确认修复

修复问题后重新运行 Pipeline:

# 通过 CLI
ordine run <PIPELINE_ID> -i <INPUT_PATH>

# 通过 API
curl -X POST http://localhost:9433/api/pipelines/<PIPELINE_ID>/run \
  -H "Content-Type: application/json" \
  -d '{ "inputPath": "<INPUT_PATH>" }'

验证新 Job 状态为 completed 且无 error 级别 Trace。

Signals

GitHub stars
28
Forks
2
Last commit
Oct 2026
Advanced
Item type
skill
Key
ordine-browse-traces
Source
github.com/forge-town/ordine