Consider the following constraints when planning or making snapshots:
-
For MemoryDB clusters, snapshot and restore are available for all supported node types.
-
During any contiguous 24-hour period, you can create no more than 20 manual snapshots per cluster.
-
MemoryDB only supports taking snapshots on the cluster level. MemoryDB doesn't support taking snapshots at the shard or node level.
-
During the snapshot process, you can't run any other API or CLI operations on the cluster.
If you delete a cluster and request a final snapshot, MemoryDB always takes the snapshot from the primary nodes. This ensures that you capture the very latest data before the cluster is deleted.
