configure-alerting-rules
About
This skill configures Prometheus Alertmanager for actionable incident alerting, setting up routing trees, receivers (like Slack and PagerDuty), and features to reduce alert fatigue. Use it when implementing proactive monitoring, integrating with on-call systems, or migrating to Prometheus-based alerting. It handles the core Alertmanager configuration for automated detection and team-based alert routing.
Quick Install
Claude Code
Recommendednpx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/configure-alerting-rulesCopy and paste this command in Claude Code to install this skill
Documentation
配置告警規則
設 Prometheus 告警規則與 Alertmanager 以得可靠可行之事件通知。
見 Extended Examples 取完整配置檔與模板。
適用時機
- 實行帶自動事件偵測之主動監控
- 依嚴重度與服務責屬路由告警至對應團隊
- 以智能分組與去重減告警疲勞
- 將監控接入值班系統(PagerDuty、Opsgenie)
- 立關鍵生產議題之升呈策
- 自舊監控系統遷至 Prometheus 告警
- 建引導應對者至解之可行告警
輸入
- 必要:欲告警之 Prometheus 指標(錯誤率、延遲、飽和)
- 必要:值班輪調與升呈策
- 選擇性:待遷之既有告警定義
- 選擇性:通知管道(Slack、email、PagerDuty)
- 選擇性:常見告警之 runbook 文件
步驟
步驟一:部署 Alertmanager
裝並配 Alertmanager 以接收自 Prometheus 之告警。
Docker Compose 部署(基本結構):
version: '3.8'
services:
alertmanager:
image: prom/alertmanager:v0.26.0
ports:
- "9093:9093"
volumes:
- ./alertmanager.yml:/etc/alertmanager/alertmanager.yml
# ... (see EXAMPLES.md for complete configuration)
基本 Alertmanager 配置(alertmanager.yml 摘):
global:
resolve_timeout: 5m
slack_api_url: 'https://hooks.slack.com/services/YOUR/SLACK/WEBHOOK'
route:
receiver: 'default-receiver'
group_by: ['alertname', 'cluster', 'service']
group_wait: 30s
group_interval: 5m
repeat_interval: 4h
routes:
- match:
severity: critical
receiver: pagerduty-critical
# ... (see EXAMPLES.md for complete routing, inhibition rules, and receivers)
配 Prometheus 用 Alertmanager(prometheus.yml):
alerting:
alertmanagers:
- static_configs:
- targets:
- alertmanager:9093
timeout: 10s
api_version: v2
預期: Alertmanager UI 於 http://localhost:9093 可達;Prometheus「Status > Alertmanagers」顯示 UP。
失敗時:
- 查 Alertmanager 日誌:
docker logs alertmanager - 驗 Prometheus 可達 Alertmanager:
curl http://alertmanager:9093/api/v2/status - 測 webhook URL:
curl -X POST <SLACK_WEBHOOK_URL> -d '{"text":"test"}' - 驗 YAML 語法:
amtool check-config alertmanager.yml
步驟二:於 Prometheus 定告警規則
作條件滿足時觸之告警規則。
建告警規則檔(/etc/prometheus/rules/alerts.yml 摘):
groups:
- name: instance_alerts
interval: 30s
rules:
- alert: InstanceDown
expr: up == 0
for: 5m
labels:
severity: critical
team: infrastructure
annotations:
summary: "Instance {{ $labels.instance }} is down"
description: "{{ $labels.instance }} has been down for >5min."
runbook_url: "https://wiki.example.com/runbooks/instance-down"
- alert: HighCPUUsage
expr: 100 - (avg by(instance) (rate(node_cpu_seconds_total{mode="idle"}[5m])) * 100) > 80
for: 10m
labels:
severity: warning
annotations:
summary: "High CPU usage on {{ $labels.instance }}"
# ... (see EXAMPLES.md for complete alerts)
告警設計良實踐:
for時長:防抖動告警。多數告警用 5-10 分鐘- 描述性註:含當前值、受影響資源、runbook 連結
- 嚴重度:critical(呼叫值班)、warning(調查)、info(知會)
- 團隊標籤:令路由至正確團隊/頻道
- Runbook 連結:每告警宜有 runbook URL
載規則入 Prometheus:
# prometheus.yml
rule_files:
- "rules/*.yml"
驗並重載:
promtool check rules /etc/prometheus/rules/alerts.yml
curl -X POST http://localhost:9090/-/reload
預期: 告警見於 Prometheus「Alerts」頁;閾值超時告警觸;Alertmanager 收觸之告警。
失敗時:
- 查 Prometheus 日誌之規則求值錯
- 以
promtool check rules驗規則語法 - 於 Prometheus UI 獨立測告警查詢
- 察告警狀態轉移:Inactive → Pending → Firing
步驟三:建通知模板
設可讀且可行之通知訊息。
建模板檔(/etc/alertmanager/templates/default.tmpl 摘):
{{ define "slack.default.title" }}
[{{ .Status | toUpper }}] {{ .GroupLabels.alertname }}
{{ end }}
{{ define "slack.default.text" }}
{{ range .Alerts }}
*Alert:* {{ .Labels.alertname }}
*Severity:* {{ .Labels.severity }}
*Summary:* {{ .Annotations.summary }}
{{ if .Annotations.runbook_url }}*Runbook:* {{ .Annotations.runbook_url }}{{ end }}
{{ end }}
{{ end }}
# ... (see EXAMPLES.md for complete email and PagerDuty templates)
於 receivers 用模板:
receivers:
- name: 'slack-custom'
slack_configs:
- channel: '#alerts'
title: '{{ template "slack.default.title" . }}'
text: '{{ template "slack.default.text" . }}'
預期: 通知格式一致、含所有相關脈絡、附 runbook 連結可行。
失敗時:
- 測模板渲染:
amtool template test --config.file=alertmanager.yml - 於 Alertmanager 日誌查模板語法錯
- 用
{{ . | json }}以調試模板數據結構
步驟四:配路由與分組
以智能路由規則優化告警遞送。
進階路由配置(摘):
route:
receiver: 'default-receiver'
group_by: ['alertname', 'cluster', 'service']
group_wait: 30s
routes:
- match:
team: platform
receiver: 'team-platform'
routes:
- match:
severity: critical
receiver: 'pagerduty-platform'
group_wait: 10s
repeat_interval: 15m
continue: true # Also send to Slack
# ... (see EXAMPLES.md for complete routing with time intervals)
分組策略:
# Group by alertname: All HighCPU alerts bundled together
group_by: ['alertname']
# Group by alertname AND cluster: Separate notifications per cluster
group_by: ['alertname', 'cluster']
預期: 告警路由至正確團隊、合邏輯分組、時序合嚴重度。
失敗時:
- 測路由:
amtool config routes test --config.file=alertmanager.yml --alertname=HighCPU --label=severity=critical - 查路由樹:
amtool config routes show --config.file=alertmanager.yml - 若告警當匹配多路由,驗
continue: true
步驟五:實行抑制與靜默
以抑制規則與暫時靜默減告警雜音。
抑制規則(抑從屬告警):
inhibit_rules:
# Cluster down suppresses all node alerts in that cluster
- source_match:
alertname: 'ClusterDown'
severity: 'critical'
target_match_re:
alertname: '(InstanceDown|HighCPU|HighMemory)'
equal: ['cluster']
# Service down suppresses latency and error alerts
- source_match:
alertname: 'ServiceDown'
target_match_re:
alertname: '(HighLatency|HighErrorRate)'
equal: ['service', 'namespace']
# ... (see EXAMPLES.md for more inhibition patterns)
程序式建靜默:
# Silence during maintenance
amtool silence add \
instance=app-server-1 \
--author="ops-team" \
--comment="Scheduled maintenance" \
--duration=2h
# List and manage silences
amtool silence query
amtool silence expire <SILENCE_ID>
預期: 抑制自動減連鎖告警;靜默於計畫維護中防通知。
失敗時:
- 以活告警測抑制邏輯
- 查 Alertmanager UI「Silences」分頁
- 驗靜默匹配器精確(標籤須完全同)
步驟六:接外部系統
連 Alertmanager 至 PagerDuty、Opsgenie、Jira 等。
PagerDuty 整合(摘):
receivers:
- name: 'pagerduty'
pagerduty_configs:
- routing_key: 'YOUR_INTEGRATION_KEY'
severity: '{{ .CommonLabels.severity }}'
description: '{{ range .Alerts.Firing }}{{ .Annotations.summary }}{{ end }}'
details:
firing: '{{ .Alerts.Firing | len }}'
alertname: '{{ .GroupLabels.alertname }}'
# ... (see EXAMPLES.md for complete integration examples)
自訂整合之 Webhook:
receivers:
- name: 'webhook-custom'
webhook_configs:
- url: 'https://your-webhook-endpoint.com/alerts'
send_resolved: true
預期: 告警於 PagerDuty 建事件、見於團隊通訊頻道、觸值班升呈。
失敗時:
- 驗 API 金鑰/令牌有效
- 查至外服之網路連通
- 以 curl 獨立測 webhook 端點
- 啟除錯模式:
--log.level=debug
驗證
- Alertmanager 成功自 Prometheus 收告警
- 告警依標籤與嚴重度路由至正確團隊
- 通知遞至 Slack、email、PagerDuty
- 告警分組適當減通知量
- 抑制規則正確抑從屬告警
- 靜默於維護窗防通知
- 通知模板含 runbook 連結與脈絡
- 重複間隔於長駐議題中防告警疲勞
- 告警清時發解決之通知
- 外部整合(PagerDuty、Opsgenie)建事件
常見陷阱
- 告警疲勞:低優告警過多令應對者忽核心者。嚴閾值、用抑制
- 缺
for時長:無for之告警於瞬峰觸。總用 5-10 分鐘窗 - 過廣之分組:以
['...']分組發個別通知。用具體標籤分組 - 無 runbook 連結:無 runbook 之告警令應對者猜。每告警需 runbook URL
- 錯嚴重度:誤標警為核令團隊脫敏。核留於急
- 遺忘之靜默:無期靜默可藏真問題。總設終時
- 單路由:所有告警至一頻道失脈絡。用團隊專屬路由
- 無抑制:斷擺時之連鎖告警生雜。實行抑制規則
相關技能
setup-prometheus-monitoring- 定供告警規則之指標與記錄規則define-slo-sli-sla- 生錯誤預算管理之 SLO 燒率告警write-incident-runbook- 建告警註中連結之 runbookbuild-grafana-dashboards- 視覺化告警觸發歷史與靜默模式
GitHub Repository
Related Skills
content-collections
MetaThis skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.
polymarket
MetaThis skill enables developers to build applications with the Polymarket prediction markets platform, including API integration for trading and market data. It also provides real-time data streaming via WebSocket to monitor live trades and market activity. Use it for implementing trading strategies or creating tools that process live market updates.
creating-opencode-plugins
MetaThis skill helps developers create OpenCode plugins that hook into 25+ event types like commands, files, and LSP operations. It provides the plugin structure, event API specifications, and implementation patterns for JavaScript/TypeScript modules. Use it when you need to intercept, monitor, or extend the OpenCode AI assistant's lifecycle with custom event-driven logic.
sglang
MetaSGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.
