Rolling Deployment
Rolling deployment is a release strategy that incrementally replaces instances running the old version of an application with instances running the new version, rather than switching everything at once.
Definition
Rolling deployment is a release strategy that incrementally replaces instances running the old version of an application with instances running the new version, rather than switching everything at once.
Overview
In a rolling deployment, a fixed pool of application instances is updated in batches: a small number of old instances are taken out of service and replaced with new ones, health-checked, and added back to the load-balanced pool before the next batch is updated. This continues until every instance is running the new version, without ever having zero healthy instances available to serve traffic. Rolling deployment is the default update strategy in Kubernetes for Deployments, where the `maxSurge` and `maxUnavailable` settings control how aggressively old pods are replaced with new ones. It requires less duplicate infrastructure than blue-green deployment, since it doesn't need two complete parallel environments, but rollback is slower and both old and new versions serve traffic simultaneously during the rollout, which requires the two versions to be compatible with each other (for example, sharing a database schema). Compared to canary deployment, which deliberately limits exposure to a small percentage of traffic while closely monitoring metrics, rolling deployment typically proceeds through the whole fleet at a fixed pace without the same emphasis on gradual, metrics-gated exposure.
Key Concepts
- Incremental replacement of old instances with new ones in batches
- Continuous availability — some healthy instances always serving traffic
- Default update strategy for Kubernetes Deployments
- Configurable batch size and surge capacity during rollout
- Lower infrastructure overhead than blue-green deployment
- Requires old and new versions to be compatible during the transition
Use Cases
Frequently Asked Questions
From the Blog
What Is Blue-Green Deployment
Blue-green deployment runs two identical production environments and switches traffic between them, giving you zero-downtime releases and instant rollback if something breaks.
Read More Cloud & CybersecurityWhat Is Canary Deployment Explained
Canary deployment releases a new version to a small slice of users first, watches key metrics, then gradually shifts all traffic over — catching problems before they hit everyone.
Read More Data ScienceHow to backtest a forecast with rolling-origin evaluation
Rolling-origin backtesting picks a cut-off, fits on history only, forecasts the full horizon, then rolls the cut-off forward and repeats. Aggregate the errors by horizon step rather than overall, because a model can be excellent one step ahead and useless at the horizon the business plans on. This article walks the mechanics and the choices inside them.
Read More Projects & Case StudiesProject: Build a REST API with Python and FastAPI
FastAPI is the fastest-growing Python web framework — and for good reason. In this hands-on project you'll build a fully functional REST API with auto-generated documentation, database persistence, and deployment on Render, all in a single afternoon.
Read More