High Availability

Koyeb provides several features to help you achieve better availability for your applications. This page outlines some of the configuration options you should evaluate if high availability is an operational requirement.

Multi-region Deployments

When configuring Koyeb Services, you can choose the regions that your application will be deployed to. Deploying to multiple regions can help you avoid disruptions caused by regional incidents.

Koyeb's global load balancer automatically routes from edge locations to the closest core region where your Service is deployed. In the event of a region-specific issue, the network will transparently route requests to other regions as required. This inter-region routing is also performed for traffic between Koyeb Services, allowing dependent systems to remain available.

Instance replicas

Within an individual region, you can choose the number of Instances of your Service to deploy. Each healthy Instance will receive a proportional share of the incoming requests for the region.

Configuring multiple Instances of your Service within a region helps performance by decreasing the load experienced by any single Instance. In terms of availability, this can be helpful in a variety of scenarios. At times, Instances may need to be moved to new servers during the course of their lifetime due to hardware or Service issues.

By deploying multiple Instances, your application can remain available within the region even when one or more Instances are moved or recreated.

If you'd like your Service to automatically scale the number of Instances according to configurable resource and traffic thresholds, take a look at the autoscaling documentation.