Available workload types
1. Service
- Long-running processes designed to handle network traffic (HTTP, gRPC, TCP, Websockets).
- Typically used for APIs, web apps, or microservices.
- Can be autoscaled based on CPU, memory, or network RPS.
2. Database
- Managed databases.
- Persistent storage enabled.
- Optimized for durability and stateful workloads.
3. Worker
- Worker is a special type of container designed to run indefinitely in the background without exposing any HTTP endpoints or acting as a web server.
- Unlike typical app containers that handle incoming requests, a worker is focused solely on executing background logic — continuously, or in response to internal events, queues, or schedules.
Choosing the right workload type
| Workload Type | Use Case |
|---|---|
| Service | Web apps, APIs, microservices |
| Database | Stateful data storage |
| Worker | Job / Task runner |
Best practices
- Use Services for stateless traffic-handling apps.
- Run Databases only when you need custom DB setups (prefer managed DBs if available).
- Ensure Workers have retry logic and error handling.