When to use PostgreSQL and when not to
As described above, PostgreSQL is well-suited for many use cases; however, there are situations where other database technologies may be better choices.
- Very large data warehouses, where queries need to be automatically distributed over multiple servers using MPP technology. Often, monolithic data collections exceeding ~200TB require a specialized solution. Such monoliths can also be redesigned into micro- or mini-services, in which case a Postgres-based solution can be considered.
- Data lakes with large data collections that rely on open table formats (OTF) stored on very cost-effective object store technologies instead of SSD and other high-speed storage models. Postgres is not designed to work with object stores in a read-only mode, but there are extensions such as DuckDB that can be used to access OTF from within Postgres.
- In-memory caching databases that support complex, high-volume, and low-latency applications. While Postgres has been used...
