domains / devops / production-resilience / release-it

release-it

Skill card

combined

The two copies overlap on the same topic (production reliability patterns) but are not interchangeable. The wondelai copy is a self-contained, detailed framework with a scoring rubric, six areas, and concrete parameter examples. The ciembor copy is a thin router that defers to external files (release-it.mini.md and release-it.md) and does not contain the procedure itself. Crowning one would hide a real alternative: the wondelai copy provides the actual steps, while the ciembor copy provides a different workflow (read mini rules first, consult full reference only when needed) and a different framing (Nygard-inspired rules).

Repositories

Other copies

Combined skill

---
name: release-it
description: 'Build production-ready systems with stability patterns: circuit breakers, bulkheads, timeouts, and retry logic. Use when the user mentions "production outage", "circuit breaker", "deployment pipeline", "chaos engineering", "retry storm", "health checks", "my service keeps crashing", "prevent cascading failures", or "make it resilient". Also trigger when designing resilient microservices, planning zero-downtime deployments, or capacity-planning for peak load. Covers stability patterns, capacity planning, deploy/release decoupling, and observability.'
license: MIT
---

# Release It! Framework

Framework for designing, deploying, and operating production-ready software. 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.

## Workflow

1. **Quick Diagnostic (Score 8/8):** Score the system by answering yes/no to 8 checks: timeouts, circuit breakers, bulkheads, zero-downtime deploy, deep health checks, correlated telemetry, load-tested past peak, failure injection. 1 point per "yes". 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; ≤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.
2. **Stability Anti-Patterns:** Walk every integration point and identify which anti-pattern it enables. Key anti-patterns: unbounded result sets, slow responses, blocked threads, retry storms, self-denial attacks. Close the specific crack rather than hardening at random.
3. **Stability Patterns:** Counter each anti-pattern with a stability pattern. 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. Fail Fast: reject requests you know will fail. Steady State: design automatic cleanup for logs, sessions, temp files. Let It Crash: a clean restart often beats limping along in an unknown state.
4. **Capacity and Availability:** Capacity is multi-dimensional. Plan for peak load, pre-scale before known events, and coordinate launches with capacity planning.
5. **Deploy/Release Decoupling:** Zero-downtime deployments. Decouple deploy from release so you can roll back without user impact.
6. **Observability:** Correlated telemetry. Deep health checks. Page on the breaker staying open, not on it opening.

## Code Applications

| Context | Guard/Pattern | Example |
|---------|---------------|---------|
| HTTP calls | Timeout + circuit breaker | Wrap all external calls with timeout + circuit breaker |
| Database queries | Result set limits | Add LIMIT; paginate all list endpoints |
| Thread pools | Isolate pools per dependency | Separate pool for payment gateway vs. search |
| 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 |
| Marketing events | Coordinate launches with capacity planning | Pre-scale before Black Friday; queue coupon redemptions |

## Conflicts

- **ciembor/agent-rules-books** instructs the agent to read `release-it.mini.md` before making decisions and to use `release-it.md` only as a deeper reference. This skill does not include those external files; the full framework is provided inline above.
- **wondelai/skills** provides the full framework inline with a scoring rubric and concrete parameter examples. The ciembor copy does not include these details and defers to external files.

## References

See [references/anti-patterns.md](references/anti-patterns.md) when triaging an outage or hardening an integration point.
See [references/stability-patterns.md](references/stability-patterns.md) when implementing a breaker or tuning thresholds.