By default the API server has 1 replica. If the pod fails, the single instance restarts by normal Kubernetes behavior, but there is downtime. If the user is upgrading, some downtime is expected.
Users have the option to configure the number of replicas using the app_replicas text box in the scst-store-values.yaml file.
By default, the database has 1 replica, and restarts with some downtime if it fails. Although the text box db_replicas exists and is configurable by the user in the scst-store-values.yaml file, VMware discourages users from configuring db_replicas because it is experimental.
The default internal database is not for use in production. For production deployments, VMware reccomends using an external database.
For the default PostgreSQL database deployment, with deploy_internal_db set to true, Velero can be used as the backup method. For information about using Velero as back up, see Backups.
