As part of its fault-tolerance policy, Soveren gateway launches several proxy instances running simultaneously when deployed with Kubernetes. When deployed with Docker Compose, you can set the number of instances.
You need a failover plan to reroute traffic back to your services, and generally to avoid making Soveren gateway a single point of failure in your perimeter. Taking into account the variety of possible environment setups, we allow you to configure the exact details of your failover plan.
Here are some tips:
Deployed with Kubernetes, Soveren gateway is designed to run three copies of the proxy simultaneously for load balancing and availability purposes. These copies don't need different Soveren tokens, as they run within a single Soveren gateway setup.
Failure of the messaging system, Apache Kafka, does not make a proxy instance fail. You may still see a gateway setup running but no stats or metrics in Soveren cloud.
Kubernetes has a built-in balancing system and manages pod lifecycles automatically, including rebalancing pods in emergency cases.