Production Readiness

Production Readiness

Overview

Code that passes all tests but logs errors only via debug-only APIs (e.g., debugPrint, console.log in debug mode) is functionally silent in production. This skill enforces observability readiness before ship — ensuring the team is not blind during beta testing and production rollout.

When to Use

• Auto-recommended for feature and architecture-change classifications.
• Activate manually when working on error handling, crash reporting, or logging infrastructure.
• Apply at /review (semantic error check) and /ship (readiness checklist).

Observability Checklist

1. Error Surface Audit (at /review)

For every catch / error-handling block in changed files:

2. Log Sink Documentation (at /ship)

Document in Work Log where production errors go:

## Observability
- Error sink: [e.g., Sentry via Logger.error(), Crashlytics, stdout → CloudWatch]
- Health check: [e.g., /health endpoint, Firebase Vitals, uptime monitor]
- Rollback signal: [e.g., error rate > 2x baseline → revert]

If the project has no production logging infrastructure, document as Known Risk:

"No production error reporting configured. Errors in catch blocks will be logged to stdout only. Risk: silent failures in release builds if stdout is not monitored."

3. Rollback Telemetry (at /ship)

The rollback plan (per engineering_guardrails.md §12.5) must answer:

• How will operators know the rollback is needed? (alert, dashboard, manual check)
• How will operators know the rollback succeeded? (error rate drops, health check passes)

Heading-Scoped Read Note

For phase-entry loading, read only:

• When to Use
• Observability Checklist

Load What This Skill Does NOT Cover, Anti-Patterns, and Interaction with Other Skills on full read or cache miss only.

What This Skill Does NOT Cover

• Tool selection: Sentry vs Crashlytics vs Datadog — team/project decides.
• Alert thresholds: When to page on-call — SRE/ops owns this.
• Dashboards: What to visualize — team preference.
• Post-deploy monitoring: This is a pre-ship readiness check, not an ops workflow.

Anti-Patterns

Interaction with Other Skills

• systematic-debugging: Complements — debugging finds bugs, production-readiness ensures they're visible when they happen again.
• verification-before-completion: Works alongside — verification checks correctness, production-readiness checks observability.
• auth-security: Security errors are especially critical to observe in production. §5.2a applies with elevated priority to auth error paths.