について
release-itスキルは、サーキットブレーカー、バルクヘッド、リトライロジックといった安定性パターンを実装することで、開発者が耐障害性の高い本番システムを構築することを支援します。フォールトトレラントなマイクロサービスの設計、ゼロダウンタイムデプロイメントの計画、障害時の連鎖的故障の防止などに活用できます。キャパシティプランニング、デプロイ/リリースの分離、オブザーバビリティを含む、本番環境対応に不可欠な要素を網羅しています。
クイックインストール
Claude Code
推奨npx skills add wondelai/skills -a claude-code/plugin add https://github.com/wondelai/skillsgit clone https://github.com/wondelai/skills.git ~/.claude/skills/release-itこのコマンドをClaude Codeにコピー&ペーストしてスキルをインストールします
ドキュメント
Release It! Framework
Framework for designing, deploying, and operating production-ready software. The software that passes QA is not the software that survives production — production is hostile, and systems must expect and handle failure at every level.
Core Principle
Every system will eventually be pushed beyond its design limits. The question is not whether failures happen, but whether your system degrades gracefully or collapses catastrophically. Production-ready software is not just correct — it is resilient, observable, and operates through partial failures without human intervention.
Scoring
Goal: 8/8. Score a production system by the Quick Diagnostic: 1 point per row answered "yes" across the 8 checks (timeouts, circuit breakers, bulkheads, zero-downtime deploy, deep health checks, correlated telemetry, load-tested past peak, failure injection). Bands: 7-8 = every integration point is bounded, isolated, observable, and deploy/release are decoupled; 4-5 = some patterns present but ≥3 diagnostic rows fail (e.g. unbounded retries, shared pools, shallow health checks); ≤2 = relies on the happy path with no breakers, no capacity model, no failure testing. Always state the current score, the failing rows, and the specific fix for each.
The Release It! Framework
Six areas that determine whether software survives contact with production:
1. Stability Anti-Patterns
Core concept: Failures propagate through integration points and cascade across system boundaries. The most dangerous patterns are not bugs in your code — they are emergent behaviors when systems interact under stress.
Why it works: These patterns recur across outages, so audit by name: walk every integration point and ask which anti-pattern it currently enables, then close that specific crack rather than hardening at random.
Key insights:
- Integration points are the number-one killer — every socket, HTTP call, or queue is a risk
- Slow responses are worse than no response: they tie up threads, exhaust pools, and propagate delay up the call chain
- Unbounded result sets turn a harmless query into an out-of-memory crash once data outgrows test assumptions
- Users generate load no test predicts — bots, retry storms, flash crowds; self-denial attacks happen when your own marketing overwhelms your infrastructure
- Blocked threads are the silent killer — deadlocks and contention show no errors until everything stops
Code applications:
| Context | Guard | Example |
|---|---|---|
| HTTP calls | Assume every remote call can fail, hang, or return garbage | Wrap all external calls with timeout + circuit breaker |
| Database queries | Enforce result set limits | Add LIMIT; paginate all list endpoints |
| Thread pools | Isolate pools per dependency | Separate pool for payment gateway vs. search |
| Marketing events | Coordinate launches with capacity planning | Pre-scale before Black Friday; queue coupon redemptions |
See references/anti-patterns.md when triaging an outage or hardening an integration point — each anti-pattern with its failure scenario and the symptom that detects it.
2. Stability Patterns
Core concept: Counter each anti-pattern with a stability pattern: circuit breakers stop cascades, bulkheads isolate blast radius, timeouts reclaim stuck resources. Together they make a system bend under load instead of breaking.
Why it works: Each pattern caps the damage one failure can do: a breaker trip converts an unbounded cascade into a fast local rejection, a bulkhead confines the outage to one pool. Treat a tripped breaker as expected output, not an incident — page on the breaker staying open, not on it opening.
Key insights:
- Circuit Breaker: three states (closed, open, half-open) — trips after threshold failures, periodically tests recovery
- Timeouts: every outbound call needs connect AND read timeouts, propagated up the call chain
- Retry with exponential backoff + jitter prevents thundering herd on recovery
- Fail Fast: reject requests you know will fail instead of wasting resources; Handshaking lets the server decline work before it's sent
- Steady State: systems accumulate cruft (logs, sessions, temp files) — design automatic cleanup
- Let It Crash: a clean restart often beats limping along in an unknown state
Code applications:
| Context | Pattern | Example |
|---|---|---|
| Service calls | Circuit Breaker | Open after 5 failures in 60s; half-open after 30s |
| Resource isolation | Bulkhead | Dedicated connection pools for critical vs. non-critical |
| Network calls | Timeout with propagation | Connect 1s, read 5s; propagate deadline downstream |
| Retries | Backoff + jitter + budget | Base 100ms, max 3 retries, 20% fleet retry budget |
| Data cleanup | Steady State | Purge sessions >24h; rotate logs at 500MB |
See references/stability-patterns.md when implementing a breaker or tuning thresholds — state-machine diagram, parameter ranges, what-counts-as-failure tables, and how to combine patterns.
3. Capacity and Availability
Core concept: Capacity is not one number — it is a multi-dimensional function of CPU, memory, network, disk I/O, connection pools, and threads. Capacity planning means knowing which resource bottlenecks first, and at what load.
Why it works: Untested systems fail at peak load — the worst possible moment. Knowing actual (not theoretical) limits lets you set realistic SLAs and scale before users hit the wall.
Key insights:
- Test taxonomy: load test (expected traffic), stress test (beyond limits), soak test (sustained, catches leaks), spike test (sudden bursts)
- Universal Scalability Law: throughput never scales linearly — contention and coherence costs cause diminishing returns
- Pool exhaustion looks identical to a database outage from the application's perspective; size pools from measured concurrency, not defaults
- "The cloud is infinitely scalable" is a myth — auto-scaling has lag, cold starts, and hard limits
Code applications:
| Context | Pattern | Example |
|---|---|---|
| Load testing | Ramp to peak, then 2x, observe degradation | Increase RPS until latency exceeds SLO |
| Connection pools | Size from measured concurrency | Set pool to P99 active connections + 20% headroom |
| Soak testing | 80% capacity for 24-72 hours | Catch memory/connection/file-handle leaks |
| Capacity model | Document bottleneck per service | "Service X is memory-bound at 2000 RPS; 4GB per instance" |
See references/capacity-planning.md when planning a load test or sizing pools — test methodologies, pool/thread tuning, and Universal Scalability Law modeling.
4. Deployment and Release
Core concept: Deployment (putting code on servers) and release (exposing it to users) are separate operations that should be decoupled — deploy without risk, release with confidence.
Why it works: Most outages are caused by changes. Decoupling lets you deploy to production, verify, and only then route traffic; if something breaks, you roll back the release, not the deployment.
Key insights:
- Zero-downtime deployment is non-negotiable: rolling, blue-green, or canary
- Feature flags dark-launch code and enable it independently of deployment
- Database migrations must be backward-compatible — old and new code run simultaneously during deploys (expand-contract)
- Immutable infrastructure: never patch a running server — build a new image, deploy, destroy the old
- Rollback must be faster than roll-forward; if rollback takes 30 minutes, you will avoid deploying
Code applications:
| Context | Pattern | Example |
|---|---|---|
| Deploys | Blue-green with health check gate | Deploy to green; smoke test; swap router |
| Progressive rollout | Canary with automated rollback | 5% traffic to canary; auto-rollback if error rate >1% |
| Feature launch | Flags with emergency off switch | Ship behind flag; enable for 10%; monitor; ramp |
| Schema changes | Expand-contract migration | Add column; write both; backfill; drop old |
See references/deployment-strategies.md when planning a release or a schema change — blue-green/canary/rolling mechanics, expand-contract migration steps, and infrastructure-as-code.
5. Health Checks and Observability
Core concept: You cannot operate what you cannot observe. Health checks, metrics, logs, and traces are the sensory organs of your system in production — a first-class design concern, not an afterthought.
Why it works: Untraced failures are invisible until a user reports them. Emit high-cardinality, structured events (not just pre-aggregated counters) so you can ask new questions of past incidents without shipping new instrumentation first.
Key insights:
- Health checks come in two flavors: shallow (process alive) and deep (dependencies reachable, resources available)
- Three pillars: structured logs (what happened), metrics (how much), distributed traces (where and how long)
- RED method for services: Rate, Errors, Duration; USE method for resources: Utilization, Saturation, Errors
- Define SLIs (measure user experience) → SLOs (targets) → SLAs (contracts), in that order
- Alert on symptoms users feel (error rate, latency), not causes (CPU); dashboards should answer "is the system healthy?" within 5 seconds
Code applications:
| Context | Pattern | Example |
|---|---|---|
| Health endpoints | Deep health check | /health reports DB, cache, queue, disk status |
| Service metrics | RED instrumentation | Rate, error rate, p50/p95/p99 latency per endpoint |
| Distributed tracing | Propagate trace context | Trace ID in headers; correlate logs across services |
| Alerting | SLO burn rate, not raw thresholds | "Error budget burning 10x" vs. "CPU > 80%" |
See references/observability.md when instrumenting a service or setting SLOs — health-check design, RED/USE metric sets, the SLI→SLO→SLA chain, and burn-rate alerting.
6. Adaptation and Chaos Engineering
Safety note: Chaos engineering experiments are design-time planning activities. The patterns below describe what to test and what to verify, not actions for an AI agent to execute autonomously. All failure injection must be performed by authorized engineers using dedicated tooling (e.g., Gremlin, Litmus, AWS FIS) with proper approvals, rollback plans, and blast radius controls in place.
Core concept: Confidence in resilience comes from testing under realistic failure conditions. Chaos engineering experiments on a system in a controlled way to build confidence it withstands turbulence.
Why it works: You cannot know how a system handles failure until it actually fails; controlled injection turns unknown-unknowns into known-knowns before they cause real outages.
Key insights:
- Define steady state first — you need a measurable baseline to detect deviation
- Every experiment has a hypothesis: "We believe that when X fails, the system will Y"
- Start small in non-production (kill one process, add latency to one call), then escalate gradually with approvals
- Minimize blast radius: canary populations, feature flags, emergency stop; production experiments require explicit authorization and instant rollback
- Automate recurring experiments; GameDay exercises test both the system and the team
- Build a culture where finding weaknesses is celebrated, not punished
Code applications:
| Context | Pattern | Example |
|---|---|---|
| Process failure | Controlled termination via chaos tooling | Kill one pod with Gremlin/Litmus; verify recovery within SLO |
| Network failure | Inject latency/partition via chaos tooling | +500ms on DB calls; verify circuit breaker trips |
| Dependency failure | Simulate downstream outage via chaos tooling | Return 503 from payment API; verify graceful degradation |
| GameDay | Scheduled team exercise | "Primary DB goes read-only at 2pm" — practice response |
See references/chaos-engineering.md when designing a failure experiment or GameDay — steady-state hypothesis, blast-radius controls, and how to grow the practice from non-prod outward.
Common Mistakes
| Mistake | Why It Fails | Fix |
|---|---|---|
| No timeouts on outbound calls | One slow dependency freezes the system | Connect and read timeouts on every external call |
| Unbounded retries | Retry storms amplify failures | Exponential backoff, jitter, fleet-wide retry budgets |
| Shared thread/connection pools | One failing dependency drains everything | Bulkhead: isolate pools per dependency |
| Shallow health checks only | Traffic routed to instances with broken dependencies | Deep health checks that verify downstream connectivity |
| Testing only the happy path | Works perfectly until the first real failure | Load, soak, and chaos test before major releases |
| Coupling deploy and release | Every deployment is all-or-nothing high risk | Feature flags, canary, blue-green |
| Alerting on causes, not symptoms | CPU alerts fire while users suffer silently | Alert on user-facing SLIs: errors, latency, availability |
| No capacity model | System falls over at 2x load | Model bottlenecks; load test to 3x expected peak |
Quick Diagnostic
Audit any production system:
| Question | If No | Action |
|---|---|---|
| Does every outbound call have a timeout? | Calls hang, blocking threads | Add connect and read timeouts everywhere |
| Are circuit breakers on critical dependencies? | One failure takes down the system | Add breakers with tuned thresholds |
| Are pools isolated per dependency? | Failures cross-contaminate | Implement bulkheads with dedicated pools |
| Can you deploy without downtime? | Deployments cause outages | Rolling, blue-green, or canary deployment |
| Do health checks verify dependencies? | Dead instances receive traffic | Deep health checks testing DB, cache, queue |
| Are logs, metrics, and traces correlated? | Debugging means manual log searches | Distributed tracing with correlated IDs |
| Have you load-tested beyond expected peak? | Unknown failure mode under real load | Test to 2-3x peak; document the breaking point |
| Do you practice failure injection? | Resilience is theoretical | Start chaos engineering with low-risk experiments |
Further Reading
For the complete methodology, war stories, and implementation details:
- "Release It! Design and Deploy Production-Ready Software" (2nd Edition) by Michael T. Nygard
About the Author
Michael T. Nygard is a software architect with 30+ years building and operating large-scale production systems handling millions of transactions per day. Release It! (2007; 2nd edition 2018) became a foundational text of the DevOps and site reliability engineering movements, arguing that architects must stay responsible for systems long after the code is written.
GitHub リポジトリ
よくある質問
release-it Skillとは何ですか?
release-it はwondelai が作成した Claude Skillです。Skillは、Claudeが必要に応じて読み込む指示とリソースをまとめ、追加の指示なしで release-it に関連するタスクを実行できるようにします。
release-it をインストールするには?
このページのインストールコマンドを使用してください。release-it をプラグインとして Claude Code に追加するか、リポジトリを skills ディレクトリにクローンし、Claudeを再起動してSkillを読み込みます。
release-it はどのカテゴリに属しますか?
release-it は メタ カテゴリに属します。
release-it は無料で利用できますか?
はい。release-it は AIMCP に掲載されており、無料でインストールできます。
関連スキル
このスキルは、Content Collections(Markdown/MDXファイルを型安全なデータコレクションに変換するTypeScriptファーストのツール)の本番環境でテストされた設定を提供します。Zodバリデーションによる型安全性を実現し、ブログ、ドキュメントサイト、コンテンツ重視のVite + Reactアプリケーション構築時にご利用ください。Viteプラグインの設定、MDXコンパイルから、デプロイ最適化、スキーマバリデーションまで、すべてを網羅しています。
このスキルは、開発者がPolymarket予測市場プラットフォームを活用したアプリケーション構築を可能にします。API統合による取引や市場データの取得に加え、WebSocketを介したリアルタイムデータストリーミングにより、ライブ取引や市場活動を監視できます。取引戦略の実装や、ライブ市場更新を処理するツールの作成にご利用ください。
このスキルは、開発者がコマンド、ファイル、LSP操作など25種類以上のイベントタイプにフックするOpenCodeプラグインを作成することを支援します。JavaScript/TypeScriptモジュール向けに、プラグイン構造、イベントAPI仕様、および実装パターンを提供します。カスタムイベント駆動ロジックでOpenCode AIアシスタントのライフサイクルをインターセプト、監視、または拡張する必要がある場合にご利用ください。
SGLangは、高性能なLLMサービングフレームワークであり、RadixAttentionプレフィックスキャッシュを活用したJSON、正規表現、エージェントワークフロー向けの高速で構造化された生成を特長とします。特にプレフィックスが繰り返されるタスクにおいて、大幅に高速な推論を実現し、複雑な構造化出力やマルチターン対話に最適です。制約付きデコードが必要な場合や、広範なプレフィックス共有を伴うアプリケーションを構築する場合は、vLLMなどの代替案ではなくSGLangを選択してください。
