Roadmap: Production, Security, and Scaling

Production Hardening

InteropNimbus is currently running in a single-node Docker Compose setup. The roadmap for production hardening includes several key areas.

Content Security Policy

Adding strict CSP headers to the Nginx configuration:

• default-src 'self' — only load resources from the same origin
• script-src 'self' — prevent inline scripts and third-party script injection
• connect-src — whitelist only the Keycloak and API endpoints
• frame-ancestors 'none' — prevent clickjacking

Rate Limiting

Traefik middleware for rate limiting API requests:

• Per-IP rate limits on the authentication endpoints
• Separate limits for API calls vs. static asset requests
• Graduated response: slow down before blocking

Centralized Logging

Forward application and Keycloak audit logs to a centralized system:

• Loki for log aggregation (already running for other services)
• Structured JSON logs from Nginx access logs
• Keycloak authentication events for security monitoring

Scaling Path

Kubernetes Migration

The current Docker Compose setup works for a single node, but the architecture is designed to be Kubernetes-ready:

• Stateless frontend — the Nginx container can scale horizontally with no shared state
• External auth — Keycloak runs independently, so frontend replicas share the same auth provider
• Health checks — already configured for K8s liveness and readiness probes

The migration path follows the same pattern used for FhirHub's K3s deployment (covered in that blog series).

Monitoring

Planned monitoring stack:

• Prometheus — metrics collection from Traefik, Keycloak, and application endpoints
• Grafana — dashboards for request rates, error rates, and authentication metrics
• Alerting — PagerDuty or Slack notifications for SLA breaches

Traefik already exposes Prometheus metrics:

metrics:
  prometheus:
    entryPoint: metrics
    addEntryPointsLabels: true
    addServicesLabels: true
    addRoutersLabels: true

Feature Roadmap

Near-Term

• Channel management — start, stop, and restart Mirth channels from the dashboard
• Message search — full-text search across message content
• FHIR write operations — create and update FHIR resources

Medium-Term

• CI/CD pipeline — automated builds, tests, and deployments on push
• E2E testing — Playwright tests against mock APIs
• Dark mode — Tailwind's dark mode support is already configured

Long-Term

• Multi-tenant — support multiple Mirth Connect instances per organization
• Webhook notifications — alert on channel errors or SLA breaches
• Plugin architecture — extensible dashboard widgets for custom integrations