agentsclimarketplace

Observability strategy designer

Skill findscripter/everything-skills/02-engineering/observability-strategy-designer

类书式 AI Agent 技能大典 · 精选/中文化/互见成网的 500+ 开源技能,可作为 Claude Code 插件市场一键安装。A curated, cross-referenced encyclopedia of 500+ open-source agent skills.

Install
npx -y skills add findscripter/everything-skills --skill observability-strategy-designer

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

当为新服务接入可观测性、治理过于嘈杂的告警,或上量前建立 SLO 体系时使用;做指标/日志/追踪三支柱设计,产出 SLI/SLO、错误预算、多窗口燃尽率告警、Grafana 仪表盘与运行手册;不适用于纯业务数据分析或一次性故障排查。触发词:SLO、告警降噪、可观测性

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

7.5 KB, ~2.2k tokens by cl100k_base, as published. Nobody here has run it

何时使用

适用场景:

  • 为新服务接入可观测性,需要从零规划指标、日志、追踪三支柱。
  • 上生产量级前建立 SLI/SLO 体系,并配套错误预算与燃尽率告警。
  • 现有告警过于嘈杂(误报多、疲劳严重),需要降噪与覆盖度治理。
  • 设计角色化(SRE / 开发 / 高管 / 运维)的 Grafana 仪表盘与运行手册。

不该用的边界:

  • 纯业务数据分析、BI 报表(与系统可靠性无关)。
  • 一次性线上故障的临场排查(本技能产出的是体系与配置,不替代实时 oncall 操作)。
  • 仅需引入某个监控工具的安装部署,不涉及策略设计。

步骤

  1. 明确服务画像:类型(api / web / database / queue / batch / ml)、关键级别(critical / high / medium / low)、是否面向用户、依赖与合规要求,整理成服务定义 JSON。
  2. 设计 SLI/SLO:按服务类型选 SLI,按关键级别定目标,计算错误预算,生成多窗口燃尽率告警与 SLA 建议。
  3. 治理告警:分析现有告警的噪声、重复、覆盖缺口与阈值合理性,输出优化后的配置。每条告警必须可执行、对症(alert on symptoms, not causes)。
  4. 生成仪表盘:按目标角色生成覆盖黄金信号 / RED / USE 的 Grafana 面板,含模板变量与下钻路径。
  5. 落地与迭代:以 IaC/GitOps 管理配置,按 MTTD/MTTR、告警精度、SLO 达成率持续调优。

指令

环境:Python 3.7+,仅用标准库,无外部依赖。脚本位于源技能 scripts/ 目录。

# 1. 生成 SLI/SLO 框架(从服务定义文件)
python3 scripts/slo_designer.py --input assets/sample_service_api.json --output slo_framework.json
# 或用命令行参数 + 仅看摘要
python3 scripts/slo_designer.py --service-type api --criticality critical --user-facing true --service-name payment-service --summary-only

# 2. 告警优化(先只分析,再生成优化配置 / HTML 报告)
python3 scripts/alert_optimizer.py --input assets/sample_alerts.json --analyze-only
python3 scripts/alert_optimizer.py --input assets/sample_alerts.json --output optimized_alerts.json
python3 scripts/alert_optimizer.py --input assets/sample_alerts.json --report alert_analysis.html --format html

# 3. 生成仪表盘(按角色 + Grafana 兼容 JSON + 文档)
python3 scripts/dashboard_generator.py --service-type web --name "Customer Portal" --role sre --output portal_dashboard.json --doc-output portal_docs.md
python3 scripts/dashboard_generator.py --input assets/sample_service_api.json --output dashboard.json --format grafana

关键约束(来自源脚本,勿擅改):

按关键级别的 SLO 目标:

  • critical:可用性 99.99%,P95 < 100ms,P99 < 500ms,错误率 < 0.1%
  • high:99.9%,P95 < 200ms,P99 < 1000ms,错误率 < 0.5%
  • medium:99.5%,P95 < 500ms,P99 < 2000ms,错误率 < 1%
  • low:99%,P95 < 1000ms,P99 < 5000ms,错误率 < 2%

多窗口燃尽率告警(短窗 / 长窗 / 燃尽率 / 预算消耗):

  • 5m / 1h / 14.4 / 2%(页级紧急)
  • 30m / 6h / 6 / 5%
  • 2h / 1d / 3 / 10%
  • 6h / 3d / 1 / 10%(票级提醒)

各服务类型推荐 SLI:

  • api:availability、latency、error_rate、throughput
  • web:availability、latency、error_rate、page_load_time
  • database:availability、query_latency、connection_success_rate、replication_lag
  • queue:availability、message_processing_time、queue_depth、message_loss_rate
  • batch:job_success_rate、job_duration、data_freshness、resource_utilization
  • ml:model_accuracy、prediction_latency、training_success_rate、feature_freshness

仪表盘角色侧重:sre(可用性/延迟/错误/资源)、developer(延迟/错误/吞吐/业务)、executive(可用性/业务/用户体验)、ops(资源/容量/告警/部署)。仪表盘认知负荷上限 7±2 个面板/屏;红=严重、琥珀=警告、绿=健康;并在面板上叠加 SLO 目标参考线。

示例

服务定义 JSON(slo_designer 输入,必填 name/type/criticality):

{
  "name": "payment-service",
  "type": "api",
  "criticality": "critical",
  "user_facing": true,
  "dependencies": [
    { "name": "user-service", "type": "api", "criticality": "high" }
  ],
  "custom_slos": {
    "availability_target": 0.9995,
    "latency_p95_target_ms": 150,
    "error_rate_target": 0.002
  }
}

告警配置 JSON(alert_optimizer 输入,historical_data 可选但能提升分析质量):

{
  "alerts": [{
    "alert": "HighLatency",
    "expr": "histogram_quantile(0.95, rate(http_request_duration_seconds_bucket[5m])) > 0.5",
    "for": "5m",
    "labels": { "severity": "warning", "service": "payment-service" },
    "annotations": { "runbook_url": "https://runbooks.company.com/high-latency" },
    "historical_data": { "fires_per_day": 2.5, "false_positive_rate": 0.15 }
  }],
  "services": [{ "name": "payment-service", "criticality": "critical" }]
}

生成的 Prometheus 告警规则模板:

- alert: {{ service_name }}_HighLatency
  expr: histogram_quantile(0.95, rate(http_request_duration_seconds_bucket{service="{{ service_name }}"}[5m])) > {{ latency_threshold }}
  for: 5m
  labels:
    severity: warning
    service: "{{ service_name }}"

注意事项

  • 每个服务先从 1-2 个 SLO 起步并迭代;SLI 要直接反映用户体验,目标按用户需求而非技术上限设定。
  • 告警务必可执行、对症告警而非告警根因;用滞回(firing/resolving 不同阈值)、依赖抑制、分组合并来防疲劳;追求高精度(少误报)优先于高召回。
  • 高基数指标是成本与性能杀手,需检测与治理;指标/日志/追踪分别用分层保留与采样(含尾部采样)控成本。
  • 配置一律纳入版本控制(IaC/GitOps),并对告警规则用历史数据回归校验。
  • 用 MTTD、MTTR、告警精度、SLO 达成率衡量体系有效性并持续调优。

互见

  • 研发/observability 同域可参考:分布式追踪、Prometheus 配置、Grafana 仪表盘、SLO 实施等相邻技能。
  • 上游产物(服务定义 JSON)可来自架构设计类技能;下游可对接事件管理(PagerDuty/告警路由)与 CI/CD 流水线监控。

采编自 alirezarezvani/claude-skills(MIT 许可)。

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.