What Is a Rolling Deployment? Incremental Updates Explained
Rolling deployment replaces instances one at a time, keeping the app available throughout. Here's how Fly.io handles it.
A rolling deployment replaces old application instances with new ones incrementally — one at a time, or in small batches. The app stays available throughout because at least some instances are always running.
How Fly.io Implements It
When you run fly deploy, Fly terminates old VMs and launches new ones in a controlled sequence. Fly waits for each new VM to pass health checks before proceeding to the next. The default strategy is rolling with a minHealthy count.
Rolling vs Blue-Green
Rolling deployments use fewer resources (no duplicate environment) but take longer and carry more risk of partial failure mid-rollout. Blue-green is safer for critical services; rolling is fine for less risky changes.
Configuring Rollbacks on Fly
If a rolling deployment fails health checks, Fly automatically stops and keeps the old VMs running. You can also manually fly releases to see history and fly deploy --image <previous> to roll back.