Facility Service — Deployment Topology

Status: populated Owner: TBD Last updated: 2026-04-17 Companion: 17 Technology Stack · SERVICE_TEMPLATE §12

1. Runtime

Attribute	Value
Language / runtime	Node 22 / TypeScript 5.x
Framework	NestJS 11
Container base	node:22-alpine multi-stage
Process model	Single-process per pod; cluster mode disabled (Kubernetes handles replication)
Port	8080 HTTP, 8081 health/metrics

2. Dependencies

Dep	Type	Purpose
PostgreSQL 16	Required	Primary store
Redis 7	Required	Cache, rate limits, idempotency
NATS JetStream	Required	Event stream
Kong	Required	Edge + auth
access-policy	Required	Authorisation
identity-service	Required	JWKS
audit-service	Soft	Audit sink
interop-service	Soft	FHIR projection

3. Replication & Scaling

Env	Replicas	Policy
dev	1	Manual
staging	2	HPA CPU ≥ 60 %, p95 latency
production	3–10	HPA CPU ≥ 70 % or request rate ≥ 800 rps/pod; min=3

Redis: ElastiCache cluster mode, 3 shards. Postgres: Aurora-compatible, 1 writer + 2 readers.

4. Regions

Region	Tenants	Notes
af-central-1	AFG MoPH + AFG private	Local DR in-region
ae-central-1	UAE DoH + UAE private
eu-central-1	Future EU pilots

Per-region isolation; cross-region reads disabled except by platform-admin with audit.

5. Environment Variables

Var	Required	Purpose
NODE_ENV	✓	development|staging|production
PORT	✓	Defaults 8080
DATABASE_URL	✓	Postgres DSN
REDIS_URL	✓	Redis DSN
NATS_URL	✓	NATS JetStream
JWKS_URI	✓	identity-service JWKS
ACCESS_POLICY_URL	✓	Access-policy internal URL
LICENSING_URL	✓	Licensing internal URL
OTEL_EXPORTER_OTLP_ENDPOINT	✓	Collector
MODULE_KEY	✓	ehr.facility
FACILITY_CONTEXT_CACHE_TTL_SECONDS	—	Default 300

6. Health & Probes

Probe	Path	Cadence
Liveness	/healthz	10s
Readiness	/readyz (db+redis+nats ping)	5s
Metrics	:8081/metrics	Prometheus scrape 15s

7. Deploy Process

• Canary 5% / 30min; auto-rollback on SLO breach.
• Blue/green for schema migrations via drizzle + expand-contract pattern.
• Feature flags for FHIR projection and edge snapshot.