#monitoring#alerting#metrics#operations
监控与告警系统
建立完整的监控和告警系统，包括指标收集、告警规则和通知渠道。
AI 摘要
这页重点
核心结论
建立完整的监控和告警系统，包括指标收集、告警规则和通知渠道。
适用主题
运维管理
高频关键词
monitoring / alerting / metrics / operations
可信信号
最后更新 2026-03-11
监控与告警系统
本文介绍 OpenClaw 的监控和告警系统。
监控指标
系统指标
| 指标 | 说明 |
|---|---|
| cpu_usage | CPU 使用率 |
| memory_usage | 内存使用率 |
| disk_usage | 磁盘使用率 |
| network_io | 网络 I/O |
应用指标
| 指标 | 说明 |
|---|---|
| requests_total | 总请求数 |
| requests_duration | 请求响应时间 |
| active_sessions | 活跃会话数 |
| errors_total | 错误总数 |
Prometheus 配置
指标端点
# prometheus.yml
scrape_configs:
- job_name: 'openclaw'
static_configs:
- targets: ['localhost:18789']
metrics_path: '/metrics'
自定义指标
from prometheus_client import Counter, Histogram
requests_total = Counter(
'openclaw_requests_total',
'Total requests',
['channel', 'status']
)
request_duration = Histogram(
'openclaw_request_duration_seconds',
'Request duration',
['channel']
)
Grafana 仪表板
仪表板配置
{
"dashboard": {
"title": "OpenClaw 监控",
"panels": [
{
"title": "请求速率",
"targets": [
{
"expr": "rate(openclaw_requests_total[5m])"
}
]
},
{
"title": "响应时间 P95",
"targets": [
{
"expr": "histogram_quantile(0.95, openclaw_request_duration_bucket)"
}
]
}
]
}
}
告警配置
告警规则
groups:
- name: openclaw
rules:
- alert: HighErrorRate
expr: rate(openclaw_errors_total[5m]) > 0.1
for: 5m
labels:
severity: critical
annotations:
summary: "错误率过高"
- alert: HighMemory
expr: openclaw_memory_usage > 0.9
for: 5m
labels:
severity: warning
annotations:
summary: "内存使用率过高"
通知渠道
route:
receiver: 'default'
routes:
- match:
severity: critical
receiver: 'critical'
receivers:
- name: 'default'
email:
to: 'admin@example.com'
- name: 'critical'
slack:
webhook_url: '${SLACK_WEBHOOK}'