High availability
Achieving a high-availability (HA) setup for the GitGuardian application requires configuring several components.
Kubernetes cluster
The foundation of a high-availability setup is a robust Kubernetes cluster.
Existing cluster
If you are using an existing cluster deployed via either KOTS or Helm, ensure it is configured for high availability. Kubernetes provides support for high availability out-of-the-box, as long as the pods are distributed across multiple nodes.
Embedded cluster
High Availability (HA) on an embedded cluster isn't recommended nor supported by GitGuardian. For HA setups, leveraging a managed existing cluster is advised.
Rolling Updates
Kubernetes natively supports rolling updates, which allows for seamless application updates without downtime. This feature is common to all installation types, ensuring that updates can be applied smoothly while maintaining high availability.
Deployments & Pods
Pod configurations are optimized to distribute across nodes evenly, minimizing the impact of a node failure. No additional configuration is required on your part.
Databases & Datastores
Both Redis and PostgreSQL require HA configurations. The embedded versions of these services don't support high availability, necessitating the use of external, HA-configured versions.
ClickHouse
ClickHouse runs as a single-node topology, which is not highly available: a node restart, upgrade, or failure takes it down until it recovers.
