Changes to the data store write through to an off-site continuous incremental backup.
The data store itself is fully redundant, each piece of information is stored multiple times, separately. Therefore, this continuous backup is the safety net in case of a required disaster recovery.
Prepared for a worst-case scenario
Possible reasons for such a recovery are:
- damage caused by human error (accidental overwrites or deletes)
- damage caused by unpermitted access, bad actors
- damage caused by a technical software bug breaking the data cluster
- hardware failure at scale, host center burning down
Cross-cluster replication
Have a read-only always up-to-date copy of your main cluster. Scenarios for this feature are:
- have a running (read-only) version of your index in case of an outage
- geographically distributed copy for faster queries in other locations
Note that this is not a backup solution, since undesired writes to the main data cluster write through to this copy.
<- Back to SearchCluster page.