| Dependency | Minimum version | Notes |
|---|
| Docker | 24.x | Required for all deployment methods. |
| Docker Compose | 2.x | Required for single-machine deployments. |
| PostgreSQL | 15.x | Primary datastore and job queue. Can be an external managed instance (RDS, Cloud SQL, etc.). |
| S3-compatible storage | — | Optional — for large job results/artefacts (AWS S3, MinIO, Cloudflare R2, or equivalent). |
| Resource | Minimum |
|---|
| CPU | 2 cores |
| RAM | 4 GB |
| Disk | 20 GB |
| Resource | Recommended |
|---|
| CPU | 4 cores |
| RAM | 8 GB |
| Disk | 50 GB SSD |
Cluster sizing depends on workflow volume and concurrency requirements. As a starting point:
- API server: 2 × (1 CPU, 512 MB RAM)
- Worker pool: scale horizontally based on concurrent run targets; each worker uses ~256 MB RAM at idle.
- PostgreSQL: managed instance recommended (RDS, Cloud SQL); also serves as the job queue, so size it for your concurrency.
- The API server must be reachable from your users’ browsers and from worker nodes.
- Workers need outbound internet access if workflows call external APIs.
- All internal component communication can stay on a private network.
