DLR Processor — Service Readiness

Status: populated Owner: Platform Engineering Last updated: 2026-04-18

1. Definition of Ready (Before Sprint Start)

• Domain model documented and reviewed
• Event schemas published to schema registry
• Data model migrations written and reviewed
• Downstream contracts agreed with billing-service and webhook-dispatcher teams
• NATS stream SMS_DLR provisioned in all environments
• Test data seeded in dev environment

2. Definition of Done (Before Merge)

• All unit and integration tests pass
• Coverage ≥ 80%
• Pact contracts verified
• npm audit passes (no CRITICAL/HIGH)
• Trivy image scan clean
• PR reviewed by ≥ 1 engineer
• Flyway migrations tested on clean DB
• Feature flag added if partially complete

3. Production Readiness Checklist

Code Quality

• No console.log or debug statements
• No hardcoded secrets or configuration
• Error handling comprehensive at all layers
• Graceful shutdown implemented (SIGTERM → drain NATS consumer → close PG pool)

Observability

• All key metrics instrumented
• Structured log events for all processing paths
• OTLP trace spans cover full pipeline
• Grafana dashboard reviewed and approved

Operations

• Runbooks written for all FM-DLR-* failure modes
• Alert rules reviewed with SRE team
• On-call rotation informed of new service
• Deployment documented and tested in staging

Security

• Security review completed
• All secrets in Vault; none in source code or environment at build time
• NetworkPolicy applied and verified
• restricted Pod Security Admission profile applied

4. Launch Phases

Phase	Criteria	Rollout
Alpha	Internal test traffic only	1 replica, staging
Beta	Canary 5% of production DLR traffic	2 replicas, production
GA	Full production traffic	3 replicas + HPA