Replacing GCP Secondary Backup with a Cost-Effective Kubernetes Operator Solution
Optimizing Database Backups: Cutting Costs, Increasing Flexibility, and Enhancing Observability with Kubernetes
We always enable the High Availability (HA) to increase reliability when using CloudSQL, which is the first choice of most Google Cloud users for database. However, it is only a tradeoff to reach a balance between cost and availability, because the default HA function works only region-wise, but does not support cross-region failover. To be more specific, once an instance in certain zone in certain region fails, HA will automatically switch to a standby instance in another zone of the same region; but when the entire region fails, HA doesn’t support to switch to another region, and you lose access to the database.
For higher reliability, there are two ready-to-use methods, cross-region replication and secondary backup.
Cross-region replication achieves asynchronous data synchronization by creating off-site replicas, but it is with the risk of data loss. It is so expensive a solution that we need to wave goodbye to.
Secondary backup backs up data regularly to other regions by enabling off-site backup. As our initial choice, it has many advantages.
