Database Service

Database Service runs PostgreSQL for you. You get a database, not a server: there is no instance to size, patch or back up by hand, and no cluster to operate.

It is the transactional half of the platform. It is built for the reads and writes an application makes while it is serving users, where each query touches a few rows and has to return quickly. Data Lake is the analytical half, built for queries that scan a great deal of data and can take their time. Sending analytical work to a transactional database is the usual way to make an application slow, which is why the two are separate products.

Where a database lives

The platform runs in several regions. Each region is a Kubernetes cluster, and a database lives in exactly one of them: the one you chose when you created it.

That choice is worth making deliberately. It decides how far your queries travel, and it cannot be changed afterwards without creating a new database and moving the data.

What you can do with one

• Create a database on a chosen region and PostgreSQL version, covered in Databases.
• Open it to the internet from named address ranges, covered in Connecting.
• Give each application its own login with its own permissions, covered in App users.
• Create schemas and design tables from the console, covered in Schemas and tables.
• Run SQL against it without leaving the browser, covered in Querying.

If you are starting from nothing, the quickstart goes from no database to a connected application in a few minutes.