Production Readiness Review Checklist Builder

Create a launch-readiness checklist for new services, features, or infrastructure changes before they go to production.

Prompt Template

You are a staff engineer running a production readiness review. Build a checklist and risk summary for the launch below.

**Project/feature:** [name]
**What is shipping:** [short description]
**Architecture summary:** [services, queues, databases, third-party deps]
**Traffic expectation:** [expected usage or load]
**User impact if it fails:** [what breaks]
**Rollout plan:** [big bang / canary / phased / feature flag]
**Observability today:** [logs, metrics, traces, alerts]
**Security/privacy concerns:** [PII, auth, permissions, compliance]
**Recovery options:** [rollback, kill switch, restore plan]
**Known risks:** [list]

Deliver:
1. Launch risk summary with red/yellow/green rating
2. Production readiness checklist grouped into architecture, testing, observability, security, data, support, and rollback
3. Missing evidence list, what must be proven before go-live
4. Launch-day runbook with owners and checkpoints
5. Go / no-go recommendation with rationale

Be strict and practical. Call out anything that would make an experienced engineer pause the launch.

Example Output

Launch Risk Summary

Rating: Yellow

The feature is functionally ready, but the team lacks saturation alerts for the new queue and has not tested rollback with live webhook traffic. Those two gaps make a full rollout risky.

Checklist Highlights

Observability

- [ ] Dashboard includes request rate, error rate, queue depth, and p95 latency

- [ ] Alert thresholds tested in staging or via replay

- [ ] Correlation IDs flow through webhook, API, and worker logs

Rollback

- [ ] Feature flag can disable writes without breaking reads

- [ ] Data backfill script has been dry-run on recent production-like data

- [ ] On-call engineer and product owner approve rollback trigger conditions

Go / No-Go

No-Go until queue-depth alerting and rollback rehearsal are complete.

Tips for Best Results

  • 💡A production readiness review should require evidence, not just verbal confidence
  • 💡If rollback depends on manual heroics, the launch is not ready
  • 💡Checklist sections should map to ownership across engineering, product, and support
  • 💡Use the same template repeatedly so teams learn what good looks like