Skip to content

README.md

Latest commit

 

History

History
17 lines (14 loc) · 1.13 KB

README.md

File metadata and controls

17 lines (14 loc) · 1.13 KB
Raw

Examples

serve() returns a plain ASGI app, so it runs on any ASGI server and mounts into any ASGI framework. None of these frameworks are dependencies of ezgh-workflows.

  • uvicorn: uvicorn app:app, where app = serve(workflows=[...], signing_secrets=[...]).
  • Starlette: Mount("/ezgh-workflows", app=serve(...)) (starlette_mount.py).
  • FastAPI: app.mount("/ezgh-workflows", serve(...)) (fastapi_app.py). Register https://<your host>/ezgh-workflows/ as the app's endpoint.
  • AWS Lambda: wrap the ASGI app with an ASGI-to-Lambda adapter (for example Mangum: handler = Mangum(serve(...), lifespan="off")) behind a function URL or API Gateway. Set max_invocation_seconds at most the function's timeout.

max_invocation_seconds must not exceed your platform's request time limit (Lambda's function timeout, Cloud Run's request timeout, your load balancer's idle timeout). EZGH sets every invocation's deadline inside it, and the SDK ends an invocation early, checkpointed, when its deadline approaches; a platform that cuts the request first loses the step in flight.