universal-dev-standards 5.15.1 → 5.17.0

package/bundled/locales/zh-TW/skills/runbook-assistant/guide.md

@@ -0,0 +1,216 @@
+---
+source: ../../../../skills/runbook-assistant/guide.md
+source_version: 1.0.0
+translation_version: 1.0.0
+last_synced: 2026-06-02
+source_hash: 79012dab91fb
+status: current
+scope: universal
+description: |
+  引導 runbook 的建立、組織與驗證。
+  使用時機：撰寫 runbook、審查維運程序、規劃演練、評估覆蓋率。
+  關鍵字：runbook, operations, incident, on-call, drill, SOP, 運維手冊, 演練。
+---
+
+# Runbook 助手
+
+> **語言**：[English](../../../../skills/runbook-assistant/guide.md) | 繁體中文
+
+**版本**：1.0.0
+**最後更新**：2026-04-01
+**適用範圍**：Claude Code Skills
+
+---
+
+## 目的
+
+此技能協助團隊撰寫、組織與驗證維運程序及 incident response 的 runbook。內容涵蓋 5 種 runbook 類型、標準 7 段式範本、撰寫品質原則、演練流程與覆蓋率報告。
+
+---
+
+## 快速參考（YAML 壓縮版）
+
+```yaml
+# === 5 RUNBOOK TYPES 五種類型 ===
+types:
+  alert_response:
+    purpose: "Diagnose and fix a specific alert / 診斷並修復特定告警"
+    example: "api-latency-high.md"
+    trigger: "Alert fires"
+  standard_operation:
+    purpose: "Routine operational procedures / 例行維運程序"
+    example: "database-backup-restore.md"
+    trigger: "Scheduled or on-demand"
+  emergency_procedure:
+    purpose: "Major incident rapid response / 重大事件快速回應"
+    example: "full-service-outage.md"
+    trigger: "SEV-1 incident"
+  change_procedure:
+    purpose: "Planned change execution steps / 計畫變更執行步驟"
+    example: "database-migration.md"
+    trigger: "Change window"
+  troubleshooting_guide:
+    purpose: "General problem investigation / 一般問題調查"
+    example: "memory-leak.md"
+    trigger: "Ad-hoc investigation"
+
+# === STANDARD TEMPLATE 標準範本 (7 Sections) ===
+template_sections:
+  1_overview:
+    fields: ["Alert Name", "Severity (P1-P4)", "Related Services", "Last Updated", "Owner", "Last Drilled"]
+  2_symptoms:
+    content: "Observable behaviors — what users see, how system behaves / 可觀察的異常行為"
+  3_impact_assessment:
+    fields: ["Affected users scope", "Business impact (revenue, reputation)", "Upstream/downstream cascading"]
+  4_diagnostic_steps:
+    format: "Numbered steps with copy-paste commands and expected output / 編號步驟+可複製指令+預期輸出"
+    branching: "If [condition A] → Scenario A; If [condition B] → Scenario B"
+  5_fix_steps:
+    format: "Per-scenario fix with verification command after each step / 每場景修復+每步驟驗證"
+    include: ["Specific commands", "Verification check", "Estimated time"]
+  6_escalation:
+    fields: ["Contact person/team", "Channel (Slack/phone)", "Backup contact", "Time threshold"]
+  7_post_actions:
+    checklist: ["Update runbook", "Create postmortem if P1/P2", "Notify stakeholders", "Record MTTD/MTTR"]
+
+optional_sections: ["Prerequisites", "Rollback Steps", "Related Runbooks", "Changelog"]
+
+# === WRITING QUALITY 撰寫品質 (6 Principles) ===
+six_principles:
+  reproducible:
+    rule: "Steps specific enough to copy-paste and execute / 步驟具體到可直接複製執行"
+    good: "kubectl rollout restart deployment/payment-service -n production"
+    bad: "Restart the payment service"
+  unambiguous:
+    rule: "Each step has exactly one interpretation / 每個步驟只有一種解讀"
+    good: "Restart the payment-service Pod in the production namespace"
+    bad: "Restart the service"
+  decision_points:
+    rule: "Branch conditions are explicit with clear criteria / 分支條件明確有具體判斷標準"
+    good: "If CPU > 90%, execute Plan A; otherwise Plan B"
+    bad: "If the server seems overloaded..."
+  rollback:
+    rule: "Failed fix steps have a fallback plan / 失敗的修復步驟有回退方案"
+  verification:
+    rule: "Each fix step has success confirmation / 每個修復步驟有成功驗證"
+    good: "Run curl health-check and verify HTTP 200"
+    bad: "Check that it's working"
+  time_bounded:
+    rule: "Expected completion time is stated / 標明預期完成時間"
+    good: "This step typically takes 2-5 minutes"
+
+anti_patterns:
+  - {bad: "Contact the team", fix: "Name specific people and channels / 指名具體人員和頻道"}
+  - {bad: "Check the logs", fix: "Provide exact log query / 提供精確的日誌查詢"}
+  - {bad: "If needed, scale up", fix: "Define threshold and target / 定義閾值和目標"}
+  - {bad: "Outdated commands", fix: "Regular review cycle / 定期審查週期"}
+
+# === DRILL PROCESS 演練機制 ===
+drill:
+  schedule:
+    highest: {condition: "P1 alert runbooks", frequency: "Monthly / 每月"}
+    high: {condition: "P2 alert runbooks", frequency: "Quarterly / 每季"}
+    medium: {condition: "Emergency procedures", frequency: "Quarterly / 每季"}
+    low: {condition: "Other runbooks", frequency: "Bi-annually / 每半年"}
+  recording:
+    fields: ["Date", "Participants", "Result (Pass/Partial/Fail)", "Estimated vs Actual time", "Issues found", "Updates required"]
+  failure_handling:
+    1: "Immediately flag runbook as needing update / 立即標記需更新"
+    2: "Create action item with owner and deadline / 建立待辦事項"
+    3: "Re-drill after update / 更新後重新演練"
+    4: "Not valid until re-drill passes / 重新演練通過前視為無效"
+
+# === REVIEW CYCLES 審查週期 ===
+review_cycles:
+  alert_response: "Quarterly (3 months) / 每季"
+  emergency_procedure: "Monthly / 每月"
+  standard_operation: "Bi-annually (6 months) / 每半年"
+  change_procedure: "After each use / 每次使用後"
+  troubleshooting_guide: "Bi-annually (6 months) / 每半年"
+
+staleness: "Runbook is stale when Last Updated exceeds review cycle or not drilled within drill cycle / 超過審查週期或演練週期即為過時"
+
+# === COVERAGE REPORT 覆蓋率報告 ===
+coverage:
+  target:
+    P1_P2: "100% alerts must have runbooks / P1/P2 告警必須 100% 有手冊"
+    P3_P4: "> 80% alerts should have runbooks / P3/P4 告警應 > 80% 有手冊"
+  report_format:
+    columns: ["Alert Level", "Total Alerts", "With Runbook", "Coverage", "Target"]
+  action: "List alerts without runbooks for creation / 列出無手冊的告警待建立"
+
+# === DIRECTORY STRUCTURE 目錄結構 ===
+organization:
+  root: "docs/runbooks/"
+  subdirs: ["alerts/", "operations/", "emergency/", "changes/", "troubleshooting/"]
+  naming: "kebab-case, describe the PROBLEM not the solution / 描述問題而非解法"
+  index: "README.md with runbook table (name, severity, services, date)"
+```
+
+---
+
+## 與相關技能整合 / Integration with Related Skills
+
+| 技能 | 整合點 | 說明 |
+|-------|-------------------|-------------|
+| `/observability --alerting` | 告警對應 runbook 連結 | 每條告警規則都應包含 `runbook_url` 標註 |
+| `/incident` | 事後檢討（postmortem）對應 runbook 流程 | postmortem 之後：更新既有或為新的失效模式建立新 runbook |
+| `/slo` | 燃燒率（burn rate）告警 | SLO 燃燒率告警應連結到對應的 runbook |
+
+### 告警連結手冊 / Alert-to-Runbook Linking
+
+```yaml
+# In alert rule definition
+annotations:
+  runbook_url: "https://wiki.example.com/runbooks/alerts/api-latency-high"
+```
+
+### 事後檢討到手冊流程 / Postmortem-to-Runbook Flow
+
+- 事件揭露了 runbook 的缺口 → 更新該 runbook
+- 事件是全新的失效模式 → 建立新的 runbook
+- 將 postmortem 連結到 runbook 的 changelog
+
+---
+
+## 設定偵測
+
+### 偵測順序
+
+1. 檢查 `CONTRIBUTING.md` 是否有「Disabled Skills」段落
+2. 檢查 `CONTRIBUTING.md` 是否有「Runbook Standards」段落
+3. 若未找到，**預設採用標準 runbook 實務**
+
+---
+
+## 詳細準則
+
+完整標準請參閱：
+- [Runbook Standards](../../core/runbook-standards.md)
+
+---
+
+## 相關標準
+
+- [Runbook Standards](../../core/runbook-standards.md) - 核心標準
+- [Alerting Standards](../../core/alerting-standards.md) - 告警設計與 runbook 連結
+- [Postmortem Standards](../../core/postmortem-standards.md) - 事後檢討
+- [Observability Standards](../../core/observability-standards.md) - 監控與儀表板
+- [Observability Assistant](../observability-assistant/SKILL.md) - 監控設定
+- [SLO Assistant](../slo-assistant/SKILL.md) - 基於 SLO 的告警
+
+---
+
+## 版本歷史
+
+| 版本 | 日期 | 變更 |
+|---------|------|---------|
+| 1.0.0 | 2026-04-01 | 首次發布 |
+
+---
+
+## 授權
+
+本技能以 [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) 授權發布。
+
+**來源**：[universal-dev-standards](https://github.com/AsiaOstrich/universal-dev-standards)

package/bundled/locales/zh-TW/skills/slo-assistant/guide.md

@@ -0,0 +1,188 @@
+---
+source: ../../../../skills/slo-assistant/guide.md
+source_version: 1.0.0
+translation_version: 1.0.0
+last_synced: 2026-06-02
+source_hash: 3c3397aec3fe
+status: current
+scope: universal
+description: |
+  引導 SLI/SLO/Error Budget 的定義與管理。
+  使用時機：定義 SLO、選取 SLI、計算 error budget、設定可靠性目標時。
+  關鍵字：SLI、SLO、SLA、error budget、reliability、burn rate、服務水準、錯誤預算。
+---
+
+# SLO Assistant
+
+> **語言**：[English](../../../../skills/slo-assistant/guide.md) | 繁體中文
+
+**版本**：1.0.0
+**最後更新**：2026-04-01
+**適用範圍**：Claude Code Skills
+
+---
+
+## 目的
+
+本技能協助團隊定義、實作與管理 Service Level Objectives（服務水準目標）。內容涵蓋不同服務類型的 SLI 選取、五步驟 SLO 設定法、Error Budget 計算與政策，以及可立即套用的範本。
+
+---
+
+## 快速參考（YAML 壓縮格式）
+
+```yaml
+# === KEY CONCEPTS 核心概念 ===
+definitions:
+  SLI: "Service Level Indicator — quantitative measurement of service behavior / 服務品質的量化量測"
+  SLO: "Service Level Objective — internal target for SLI over a window / SLI 的內部目標值"
+  SLA: "Service Level Agreement — external contract with penalties / 外部合約，違反有罰則"
+  error_budget: "Allowed unreliability = 1 - SLO target / 允許的不可靠度"
+
+relationship: "SLI ── measures ──► SLO ── stricter than ──► SLA"
+rule: "SLO should be stricter than SLA to provide buffer / SLO 應比 SLA 嚴格以提供緩衝"
+
+# === SLI SELECTION GUIDE SLI 選取指南 ===
+sli_by_service_type:
+  api_services:
+    availability: "Non-5xx / Total requests"
+    latency: "Requests below duration threshold / Total"
+    quality: "Non-degraded responses / Total"
+  batch_jobs:
+    freshness: "Time since last success within threshold"
+    correctness: "Correctly processed / Total records"
+    coverage: "Completed batches / Scheduled batches"
+  frontend:
+    load_performance: "LCP<2.5s, FID<100ms, CLS<0.1 / Total pages"
+    availability: "Successful loads / Total loads"
+    interaction_delay: "Actions with response < threshold / Total"
+
+sli_quick_select:
+  api: "→ Availability + Latency + Quality"
+  batch: "→ Freshness + Correctness + Coverage"
+  frontend: "→ LCP/FID/CLS + Availability"
+
+# === 5-STEP SLO METHODOLOGY 五步驟設定法 ===
+slo_methodology:
+  step_1_select_sli: "Choose 1-3 most meaningful SLIs per service / 每個服務選 1-3 個最有意義的 SLI"
+  step_2_measurement_window:
+    rolling_28d: "Recommended; smooths weekly patterns / 推薦，平滑週間波動"
+    calendar_month: "Aligns with business reporting / 對齊業務報告週期"
+  step_3_set_target:
+    "99%": "7.3h downtime/month — internal tools / 內部工具"
+    "99.5%": "3.65h — general B2B / 一般 B2B"
+    "99.9%": "43.8min — consumer-facing / 面向消費者"
+    "99.95%": "21.9min — financial, healthcare / 金融、醫療"
+    "99.99%": "4.38min — infrastructure, payments / 基礎設施"
+    tip: "Don't set higher than you can sustain / 別設得比你能維持的更高"
+  step_4_compliance_formula: |
+    SLO Compliance = (Good Events / Total Events) x 100%
+    Define: Good Events, Total Events, Window, Target
+  step_5_document: "Record using SLO Document Template / 使用 SLO 文件範本記錄"
+
+iterative_adjustment:
+  budget_underused: "Consider tightening SLO / 考慮收緊 SLO"
+  budget_exhausted_often: "Loosen SLO or invest in reliability / 放寬或投資可靠性"
+  users_unhappy_despite_meeting: "Revisit SLI selection / 重新檢視 SLI 選取"
+  team_burnout: "SLO may be too aggressive / SLO 可能過於激進"
+
+# === ERROR BUDGET 錯誤預算 ===
+error_budget:
+  formula: "Budget = 1 - SLO Target"
+  example: "99.9% over 28d → 0.1% → 40.32 min allowed downtime"
+  quick_calc: "Budget = (1 - target) x window_minutes"
+
+burn_rate_alerting:
+  fast_burn: {threshold: "2% in 1h", action: "Page (P1)", meaning: "Budget gone in ~2 days"}
+  medium_burn: {threshold: "5% in 6h", action: "Alert (P2)", meaning: "Budget gone in ~5 days"}
+  slow_burn: {threshold: "10% in 3d", action: "Ticket (P3)", meaning: "May exhaust before month end"}
+
+budget_exhaustion_policies:
+  freeze_releases: "Halt non-reliability features until recovery / 凍結非可靠性功能發布"
+  reliability_sprint: "Dedicate sprint to reliability / 專注衝刺可靠性改善"
+  enhanced_review: "Additional production-readiness review / 加強生產就緒審查"
+  lower_slo: "Reduce target after stakeholder agreement / 經利害關係人同意後降低目標"
+  rule: "Agree on policy BEFORE budget exhaustion / 在預算耗盡前就達成政策共識"
+
+# === SLO TEMPLATES 範本庫 ===
+templates:
+  api_service:
+    availability: {sli: "Non-5xx / Total", default_slo: "99.9%"}
+    latency: {sli: "P99 < threshold / Total", default_slo: "99%"}
+    error_rate: {sli: "Non-5xx+timeout / Total", default_slo: "99.9%"}
+  batch_job:
+    freshness: {sli: "Max delay < threshold", default_slo: "99.5%"}
+    correctness: {sli: "Correct / Total", default_slo: "99.99%"}
+    completion: {sli: "Completed / Scheduled", default_slo: "99.9%"}
+  frontend:
+    load_perf: {sli: "LCP<2.5s proportion", default_slo: "90%"}
+    interaction: {sli: "FID<100ms proportion", default_slo: "95%"}
+    visual: {sli: "CLS<0.1 proportion", default_slo: "95%"}
+```
+
+---
+
+## 與相關技能整合 / Integration with Related Skills
+
+| 技能 | 整合點 | 說明 |
+|-------|-------------------|-------------|
+| `/observability` | SLI 資料來源 | 三大支柱的指標餵入 SLI 計算 |
+| `/incident` | SLO 影響評估 | 在事件報告中納入已消耗與剩餘的 Error Budget |
+| `/sdd` | 規格中的 SLO | 新服務規格「應」包含 SLO 目標章節 |
+
+### 事件回應中的 SLO / SLO in Incident Response
+
+回報事件時，應包含：
+- **受影響的 SLI**：哪個指標受到衝擊
+- **已消耗的 Error Budget**：分鐘數或百分比
+- **剩餘預算**：該量測視窗內還剩多少
+
+### 規格驅動開發中的 SLO / SLO in Spec-Driven Development
+
+使用 `/sdd` 定義新服務時，新增一個 SLO 章節：
+- 附帶理由的 SLI 選取
+- 附帶論證的目標值
+- Error Budget 政策的選擇
+
+---
+
+## 設定偵測
+
+### 偵測順序
+
+1. 檢查 `CONTRIBUTING.md` 是否有「Disabled Skills」章節
+2. 檢查 `CONTRIBUTING.md` 是否有「SLO Standards」章節
+3. 若皆未找到，**預設採用標準 SLO 實務做法**
+
+---
+
+## 詳細準則
+
+完整標準請參閱：
+- [SLO Standards](../../core/slo-standards.md)
+
+---
+
+## 相關標準
+
+- [SLO Standards](../../core/slo-standards.md) - 核心標準
+- [Observability Standards](../../core/observability-standards.md) - 三大支柱框架
+- [Alerting Standards](../../core/alerting-standards.md) - 以 SLO 為基礎的告警
+- [Performance Standards](../../core/performance-standards.md) - 效能目標
+- [Observability Assistant](../observability-assistant/SKILL.md) - 監控設定
+- [Runbook Assistant](../runbook-assistant/SKILL.md) - 告警回應程序
+
+---
+
+## 版本歷史
+
+| 版本 | 日期 | 變更內容 |
+|---------|------|---------|
+| 1.0.0 | 2026-04-01 | 初始發布 |
+
+---
+
+## 授權
+
+本技能以 [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/) 授權發布。
+
+**來源**：[universal-dev-standards](https://github.com/AsiaOstrich/universal-dev-standards)