Data Retention

The Admin Service uses an in-memory data grid (IMDG) as a cache to hold its data. This allows state information to be shared between Admin Service instances, whether they are in the same Kubernetes cluster or distributed across multiple clusters. Any instance can use current data to execute commands in the most appropriate location.

This data is volatile and not guaranteed to survive pod restarts. Best efforts are made to ensure that it is duplicated on least two pods, but replication times vary. It is possible for a pod to be restarted before a replica is made. Command execution history and audit events may be incomplete and should only be used as a guide.

Data about executable commands is defined in the Kubernetes API and the deployed applications. Therefore, this data, while it is also volatile, is recovered when the pods restart.