One warm function. Multiple concurrent requests.
Easel Functions run your framework’s server routes without server management. Concurrent Functions let available warm runtimes process overlapping requests through in-function concurrency.
Make warm capacity go further
Most function platforms assign one in-flight invocation to an execution environment at a time. During concurrent traffic, they may provision additional environments even while existing requests are waiting on network I/O.
- GET/api/feed78 ms
- GET/api/notifications50 ms
- POST/api/events52 ms
- GET/api/users54 ms
In-function concurrency lets a warm runtime begin another request while one is waiting on the network. Request multiplexing keeps each request’s identity isolated on the request fork.
Request versus invocation
Easel separates the request you care about from the invocation that may serve multiple overlapping requests. Observability shows both views.
Four concurrent requests multiplexed onto one warm function invoke
Runtime behavior
Applications keep their existing request model. Easel manages in-function concurrency at the platform layer with a concurrency-capable runtime.
Framework-generated functions
SSR, Route Handlers, server functions, and middleware run as platform functions.
Named compute regions
Place server-side work in explicitly identified regions such as IAD, FRA, and GRU.
Cold-start and billing context
Separate initialization from application execution; see compute and billed duration per invocation.
Run Concurrent Functions on your next deploy
Deploy an app with server routes, then open Observability to inspect request and invocation traces.