Skip to main content
Important: We are not able to provide support or guidance for unmanaged Postgres. We now offer Fly.io Managed Postgres , our fully-managed database service that handles all aspects of running production PostgreSQL databases.
Illustration by Annie Ruygt of a balloon disguised as an elephant Postgres, officially PostgreSQL, is a powerful open source object relational database system that’s used by many popular web frameworks to persist application data. Fly Postgres (unmanaged) is a Fly app with flyctl sugar on top to help you bootstrap and manage a database cluster for your apps. It comes with most commonly used functionality (replication, failover, metrics, monitoring and daily snapshots). When you create a Fly Postgres cluster, you’re offered several preset configurations. The “High Availability” options are three-node clusters. In case of a node failure, a HA Fly Postgres cluster will automatically take the bad actor out of the picture and, if necessary, fail over leadership to the healthy node. Flyctl also offers a single-instance “Development” config. If you choose this option and the hardware it’s running on has network problems or the SSD fails, your database will go down. In the latter case, you will lose any data accumulated since the last snapshot. Fly Postgres is designed so that you can turn a single-node cluster into a high-availability one just by adding a second instance in the same region. In the single-instance “Development” config, you’ll also be given the option to activate “automatic scale to zero”. It’s not magic: it’s just how our apps work. If you have an app that connects to your database, make sure that it also scales to zero (otherwise the connection remaining open will prevent it from “going to sleep”). You can also add read-only replicas in other regions and take advantage of the Fly.io platform’s proxy features to build a Globally Distributed Postgres cluster. The Fly Postgres app is fully open source. Just fork fly-apps/postgres-flex and add whatever meets your needs. You can even update an existing cluster with your new image using fly deploy --image. One caveat is that once you fork, you won’t be able to use fly postgres commands to administer your app.