49 lines
2.4 KiB
JSON
49 lines
2.4 KiB
JSON
{
|
||
"module": "health_check",
|
||
"version": "1.0",
|
||
"purpose": "Tier 1 快速健康检查（每5分钟）。对 service_registry 中所有注册服务做三合一检查：PID存活 → 端口监听 → HTTP /health。异常时写 TODO 给 self_todo_executor 消费。正常时静默。",
|
||
"ui_location": "F Tab → 定时任务 → 健康检查(5min)",
|
||
|
||
"human_help": {
|
||
"title": "Tier 1 健康检查",
|
||
"description": [
|
||
"每 5 分钟由 Windows Task Scheduler 触发，检查所有注册服务的运行状态。",
|
||
"检查项：端口监听（netstat）+ HTTP /health 端点响应。远程服务跳过本地端口检测。",
|
||
"PID 检查仅作为辅助信号——端口+HTTP 都正常即判定为健康。",
|
||
"异常服务会写入 health_todos.jsonl，由自修复执行器消费。",
|
||
"全正常时静默运行，不输出任何噪音。"
|
||
],
|
||
"usage": [
|
||
"报告文件：gateway/temp/last_health_check.json",
|
||
"TODO 文件：gateway/temp/health_todos.jsonl",
|
||
"日志：gateway/logs/health_check_report.log",
|
||
"Dashboard 通过 /api/monitor 读取报告显示 Tier1 统计"
|
||
],
|
||
"troubleshooting": [
|
||
"如果定时任务未运行：检查 Windows Task Scheduler → agents-health-check",
|
||
"如果报告未更新：检查 C:\\Users\\hmo\\AppData\\Local\\Programs\\Python\\Python310\\python.exe 是否可用",
|
||
"如果 TODO 堆积：检查 self_todo_executor 是否正常消费"
|
||
],
|
||
"related": "依赖 service_registry.py（服务定义）; 输出给 self_todo_executor.py（修复消费）; 报告被 dashboard.py 读取"
|
||
},
|
||
|
||
"ai_spec": {
|
||
"dependencies": [
|
||
"service_registry.py — 服务定义（名称/端口/health URL/fix_script）",
|
||
"Python 3.10 (C:\\Users\\hmo\\AppData\\Local\\Programs\\Python\\Python310\\python.exe)",
|
||
"Windows Task Scheduler — 每5分钟触发"
|
||
],
|
||
"constraints": [
|
||
"PID 检查不是硬性失败条件——端口+HTTP OK 即视为健康",
|
||
"远程服务（remote:true）跳过本地 netstat 端口检测，直接用 HTTP 判断",
|
||
"静默原则：全正常时零输出；仅异常时写 TODO + 打印报告"
|
||
],
|
||
"related_files": [
|
||
"gateway/scripts/agents_health_check.py — 本脚本",
|
||
"gateway/scripts/service_registry.py — 服务注册表",
|
||
"gateway/scripts/self_todo_executor.py — TODO 消费者",
|
||
"gateway/scripts/specs/health_check.json — 本 spec"
|
||
]
|
||
}
|
||
}
|