Production Checklist

Complete this checklist before deploying Triage Warden to production.

Security Requirements

Authentication & Secrets

• Encryption key configured: Set TW_ENCRYPTION_KEY with a 32-byte base64-encoded key

  # Generate a secure key
  openssl rand -base64 32
  

• JWT secret configured: Set TW_JWT_SECRET with a strong random value

  openssl rand -hex 32
  

• Session secret configured: Set TW_SESSION_SECRET for session encryption

• Default admin password changed: Change the default admin credentials immediately after first login

• API keys use scoped permissions: Don't create API keys with * scope in production

Network Security

• TLS enabled: All traffic should use HTTPS
• TLS certificates valid: Use certificates from a trusted CA (not self-signed)
• Internal traffic encrypted: Database connections use TLS
• Firewall rules configured: Only expose necessary ports (443 for HTTPS)
• Rate limiting enabled: Protect against brute force attacks

Database Security

• PostgreSQL in production: Don't use SQLite for production workloads
• Database user has minimal permissions: Use a dedicated user, not superuser
• Database connections encrypted: Enable sslmode=require or verify-full
• Regular backups configured: Automated daily backups with tested restore procedure

Configuration Requirements

Required Environment Variables

Optional but Recommended

LLM Configuration (if using AI features)

• LLM API key configured: Set via UI or environment variable
• Rate limits configured: Prevent runaway API costs
• Model selected appropriately: Balance cost vs. capability

Infrastructure Requirements

Minimum Hardware

Database Requirements

Network Requirements

• Outbound HTTPS (443) to:
  • LLM provider (api.openai.com, api.anthropic.com)
  • Configured connectors (VirusTotal, Jira, etc.)
• Inbound HTTPS (443) from:
  • Users accessing the dashboard
  • Webhook sources (SIEM, EDR systems)

Monitoring & Observability

Health Checks

• Health endpoint accessible: GET /health returns component status
• Readiness probe configured: GET /ready for load balancer
• Liveness probe configured: GET /live for container orchestration

Metrics & Logging

• Prometheus metrics exposed: GET /metrics endpoint enabled
• Log aggregation configured: Logs shipped to central system
• Alerting rules configured: Alerts for critical failures

Recommended Alerts

Operational Readiness

Documentation

• Runbooks available: Team has access to operational runbooks
• Contact list current: On-call rotation and escalation paths defined
• Recovery procedures tested: Backup restore verified within last 30 days

Access Control

• Admin accounts audited: Remove unnecessary admin users
• API keys audited: Revoke unused or over-privileged keys
• Audit logging enabled: User actions are logged

Backup & Recovery

• Database backups automated: Daily backups with 30-day retention
• Backup encryption enabled: Backups encrypted at rest
• Recovery time objective defined: Team knows target RTO
• Recovery procedure documented: Step-by-step restore guide exists

Pre-Launch Testing

Functional Tests

• User login works with configured auth
• Incidents can be created via webhook
• Playbooks execute correctly
• Connectors authenticate successfully
• Notifications are delivered

Load Testing

• Tested with expected concurrent users
• Tested with expected webhook volume
• Response times acceptable under load

Failover Testing

• Application recovers from database restart
• Application handles LLM API failures gracefully
• Kill switch stops all automation when activated

Sign-Off

Quick Validation Commands

# Check health endpoint
curl -s https://triage.example.com/health | jq

# Verify TLS certificate
openssl s_client -connect triage.example.com:443 -servername triage.example.com

# Test database connectivity (from application)
curl -s https://triage.example.com/health/detailed | jq '.components.database'

# Verify all connectors healthy
curl -s https://triage.example.com/health/detailed | jq '.components.connectors'