[ahs1200]

Our Kubernetes operator attaches persistent volume claims to each ArangoDB pod. The k8s infrastructure automatically fulfills these claims with actual persistent volumes on most cloud providers. This gives you network volumes of the desired size. If a pod needs to be redeployed, it can come back with its data, even if it is scheduled to a different pod. If you are on an on-premise k8s cluster or want to use fast locally attached SSDs, we have a separated ArangoStorageOperator, which can fulfill the persistent volume claims with local volumes. In this case, of course, a pod cannot be redeployed to a different machine. However, the ArangoDB cluster layer handles synchronous replication and automatic failover automatically. Therefore, the fundamental problem of running stateful services in a k8s environment is essentially solved.