Skip to main content

Deployment Topology

:::info Source Sourced from services/assessment-service/DEPLOYMENT_TOPOLOGY.md in the documentation repo. :::

1. Containers​

  • assessment-api — REST API (quiz serve, scoring).
  • assessment-worker — AI grading async jobs, outbox relay.
  • assessment-ai-gen — AI question generation worker (pooled to ai-gateway).

2. Scaling​

ContainerMinMaxHPA
api320CPU>60%
worker210queue depth > 100
ai-gen18queue depth > 50

3. Resources​

api: 500m/2000m CPU, 512Mi/2Gi mem. worker: 500m/2000m, 512Mi/2Gi. ai-gen: 300m/1500m, 512Mi/2Gi.

4. Caching​

Redis: hot quiz banks (5 min TTL), AI grade cache (24h TTL keyed by response hash).

5. Storage​

  • Postgres schema assessment.
  • Answer key blobs encrypted in DB (AES-256 + tenant DEK).

6. CDN​

  • Public quiz preview images/assets via CDN with signed URLs.
  • No direct quiz content via CDN (security).

7. Regional​

Per-region deployment; quiz banks pinned to tenant homeRegion.

8. Service Mesh​

mTLS. Egress: progress, authoring, content, ai-gateway, notification.

9. Release​

Blue/green for API. Rolling for workers. AI gen worker scaled independently.

10. DR​

RPO 5 min, RTO 60 min. Quiz banks + scenarios rebuildable from event log.