Managed databases

PostgreSQL 18 and 17, MySQL 9.7 and 8.4, MongoDB 8.0 and 9.0 and Redis 8.2 and 7.2, created from the dashboard in minutes. Every database comes up with host, port, user, password and certificates ready — connections only happen over TLS.

What every database ships with

Default behaviour for any database created here. Nothing on this list has to be configured.

The engine, with nothing in between

PostgreSQL 18 on port 5432, MySQL 9.7 on 3306, MongoDB 8.0 on 27017 and Redis 8.2 on 6379. Your driver, ORM and migration tool talk to the instance exactly as they would to any install of those versions.

You pick the memory

You decide how much of your plan's RAM goes to the database, and you can change it later. The container restarts with the new limit; host, password and certificates stay the same.

One container per database

Each database runs in its own container on the platform's internal network. Nothing inside it is shared with another customer, and access from outside requires TLS.

Snapshots you can download

On every plan: take a manual snapshot whenever you want, restore any earlier one with a 30-second cooldown, and download its contents as a ZIP.

Up in minutes

Pick the engine and the memory; the dashboard hands you host, port, user, password and the TLS certificates to download. No server to provision, no CA to maintain.

Data tab in the dashboard

On plans with the data editor: tables, collections, keys and TTLs open right in the dashboard, SQL console included — no connection string to hand around, no client to install.

Minimum requirements

To create a database, your plan needs free RAM to allocate to it:

Minimum RAM required:
512MB

A database's memory comes out of your plan's RAM and stays unavailable to your applications for as long as it exists.

Other managed databases

Same dashboard, same certificates, a different engine.

Questions

Frequently asked questions

What people ask before creating their first database

What is a managed database here?

The platform creates the container, issues the CA and the TLS certificates, exposes the host and keeps the engine running. You get the credentials and look after your data only — no server to provision, no packages to patch, no certificate to renew by hand.

How do I create a database?

In the dashboard, pick the engine (PostgreSQL, MySQL, MongoDB or Redis) and how much of your plan's memory goes to it. Within minutes it is up, with host, port, user, password and certificates ready to download.

Which database should I choose?

Redis for cache, sessions and queues; MongoDB for documents that change shape; PostgreSQL for queries that span tables inside a transaction; MySQL when the project already expects MySQL. Each engine's page has the detail.

What is the backup policy?

On every plan, a database gets daily backups and manual snapshots: take one whenever you want, restore any earlier snapshot (with a 30-second cooldown between restores) and download its contents as a ZIP. There is no recovery to an arbitrary point in time — what exists is the snapshot that was taken.

Can I resize my database?

Yes. Change the allocated memory, within your plan's ceiling, or move to a plan with more RAM. The container restarts to come up with the new limit, so open connections drop for a moment; host, password and certificates stay the same.

Are there connection limits?

We don't publish a per-plan connection quota. The practical ceiling is the memory you allocated: every open connection takes memory from the container. If your application opens a connection per request, put a pool in front of it — that is what handles load without buying the database more RAM.

Ready to create your database?

PostgreSQL, MySQL, MongoDB or Redis from the dashboard, with mandatory TLS and certificates ready to download.

Sign in to get started