Cloud//edge computing//cold start

The extra latency on the first request to a serverless function when the runtime hasn't been initialized yet. First request → spin up runtime → load code → execute → respond. Subsequent requests → already warm → fast.

The extra latency on the first request to a serverless function when the runtime hasn't been initialized yet. First request → spin up runtime → load code → execute → respond. Subsequent requests → already warm → fast.

In Lambda (containers): cold starts range from 100ms to 1s+ depending on language and bundle size. The container has to boot.

In Workers (V8 isolates): cold starts are ~0ms because isolates are ultra-lightweight — no container, no VM, no OS boot. Just V8 spinning up a sandbox.

This is the main reason Workers feel faster than Lambda for the first request — V8 isolates start in microseconds, containers don't.

After idle (no requests for a while), the function goes cold again and the next request pays the startup cost