What actually happens when you click “Deploy”

You hit Deploy and your app goes live in seconds.

But instead of imagining a split system (CDN / edge / serverless / etc.), we simplify the model:

Your entire app runs inside one Firecracker VM
started with:
npm run start

Step 0 — Git push triggers deploy

GitHub sends a webhook to the deployment system.

A worker is picked from a warm pool almost instantly.

A fresh Firecracker microVM is created.

Inside it:

Boot is extremely fast because Firecracker is designed for serverless-style workloads.

The system checks your cache:

This is one of the biggest reasons repeat deploys feel much faster.

Your app is built normally:

npm run build

But with awareness that this is a Next.js project:

Still: everything stays inside the same VM.

No splitting into edge/CDN/serverless layers.

Now the VM runs your app:

npm run start

This launches the Next.js server inside the Firecracker VM.

It handles:

One process. One environment. Everything included.

Once the server is healthy, traffic is routed to this VM.

Older deployments remain unchanged and still runnable.

Rollback is just a pointer switch — no rebuild needed.

A user opens your site:

No infrastructure fragmentation — just request → app → response.

Instead of thinking:

Think:

One deployment = one Firecracker VM running npm run start

Simple, predictable, and fully isolated per deploy.