High availability is essential for any application, and PostgreSQL deployment options can increase high availability in various ways. In this article, the author reviews several popular PostgreSQL high availability deployment options, including single database instance, single primary instance with read replicas, multi-master with a coordinator, and multi-master without a coordinator. The article notes the difference between high availability and disaster recovery and emphasizes the importance of measuring high availability by defining uptime commitments outlined in service level agreements. The article also provides sample configurations for each deployment option and highlights the benefits of using solutions such as CitusData and YugabyteDB for sharding and multi-master deployments.
https://www.yugabyte.com/postgresql/postgresql-high-availability/