In a production environment, Hybrid Data Pipeline should be deployed on one or more nodes behind a load balancer. If a host machine goes down or becomes unavailable, best practices recommend that you maintain performance by replacing the failed node. For a deployment with a network load balancer, the following general steps apply.

  1. Install the server on a new node.
  2. Configure the load balancer to use the new node.
  3. Remove references and routing information for the failed node from the load balancer.
Note: If you are running a cloud load balancer such as the AWS Application Load Balancer or the Azure Application Gateway, you should consult your vendor documentation about managing compute capacity. For information specific to Hybrid Data Pipeline deployments using cloud load balancers, refer to Load balancer configuration in the Deployment Guide for details.