Recovery, backups and maintenance
Recovery
Availability-zone
In the event of an availability-zone failure, temporary traffic disruptions may occur. The application tier autoscaling group will recreate the number of instances required to handle all incoming traffic in about 15 minutes.
Region failure
The architecture outlined in this guide does not natively support multi-region operation. Enable the optional cross-region replication of backups to restore in a different region.
Recovery Time Objective (RTO) and Recovery Point Objective (RPO)
Deeploy is a mostly stateless application. The business critical data is housed in Amazon RDS for PostgreSQL and Amazon S3. A 4-hour recovery time objective (RTO) and 24 hour recovery point objective (RPO) are generally possible. To restore operations:
- Recreate the (Stateless) EKS cluster as described in this article and install the Deeploy stack as described here
- Recreate the RDS database for PostgreSQL as described in this article and restore the RDS backup as described here.
- Recreate the S3 storage as described in this article and restore previous version as described here.
Backup
We highly recommend to do daily back-ups and snapshots to ensure minimal loss and downtime for the RDS and use versioning for S3. For more details:
Maintenance
Regular updates should use rolling updates as provided by our updated Helm Chart as described in this article. We recommend to regularly check for updates and check the release notes. Updates can include changes to the infrastructure as well. Instructions and templates will be provided in that case.