Reliability
High-availability
Every instance is deployed in a cluster where, if an availability zone fails, it can instantly be scheduled on another node in an available AZ.
In case of a unavailability in multiple or all AZs in a cluster region, the EVERTRUST control plane will automatically migrate workloads to a cluster in the same cloud provider but in a different region.
| Instances that include VPN connectivity or network peering cannot be automatically migrated to another region. Manual action will be required on the customer side to implement changes. |
Disaster recovery
Backups are taken regularly on every instance database, triggered by the following policy:
-
full instance backup: every 6 hour
-
cluster snapshot: every hour
Full instance backups
Full instance backups are specific to a deployed instance, and can be restored upon customer request. They are replicated across multiple regions and cloud providers, and retention is 1 year.
Cluster snapshots
Cluster snapshots are used by EVERTRUST Cloud to ensure business continuity in case of an outage or data loss on MongoDB Atlas side. They are replicated in two regions in the same cloud provider, and retention is the following:
-
Hourly snapshots: 2 days
-
Daily snapshots: 7 days
-
Weekly snapshots: 4 weeks
-
Monthly snapshots: 12 months
Platform outages
The cloud platform service status is available at all times at https://status.evertrust.io. Incidents and maintenances will be reported through the status page above.
Additionally, the EVERTRUST Cloud platform has been designed to be able to operate even when some components are in a degraded state. If the control plane suffers from unavailability, some operations will be unavailable (upgrade, restores) but existing instances will continue to operate as normal.