How do you handle backups / disaster recovery?

By default, backups run every 24 hours and files are stored offsite for 7 days. Thus, clusters can be restored to any prior state within the past week. If you want backups to be run more frequently, contact support here: https://qbox.io/support

Here is full article on our backup system: https://qbox.io/backend/entrie...

For the sake of automated failover, we strongly recommend using 2 replicas with replica shards of all indices. By default, we set replica shards to nodes - 1 in Elasticsearch, which fully replicates data across all nodes. So even if all but 1 node fails, your data survives and the cluster can recover without manual intervention. Such scenarios are rare.

Here is a full article explaining node failover: https://qbox.io/backend/entrie...

Replica settings change when nodes are added/removed, though you can always manually set them depending on your needs. Our automated recovery process will always periodically attempt to restart nodes that stall or disconnect. Cluster restarts can also be issued manually from the dashboard.