Managed databases
Provision and operate PostgreSQL clusters through Relay — from creation to day-2 operations.
A managed database is a Relay-operated cluster: an engine, one or more nodes, provisioned into a provider and region. Relay owns its lifecycle — placement, networking, TLS, backups, metrics, users, parameters, scaling, and restore.
PostgreSQL is the engine available today. The engine catalog is backend-driven; additional engines appear as they become available in your organization.
In this section
- Provisioning — engines, regions, machine tiers, and single-node vs. high-availability.
- Topology & multi-region — node groups, replicas, and disaster-recovery layouts.
- Connecting — connection strings, TLS, and endpoints.
- Database users & roles — bounded users with show-once passwords.
- Logical databases.
- Parameters & extensions.
- Metrics & logs.
- Backups & restore.
- Scaling & storage.
- Lifecycle — pause, resume, delete.
The model
A cluster runs one or more nodes across node groups. A single-node cluster is one primary; a high-availability cluster adds a streaming replica with automatic failover. Storage, compute (machine tier), and disk characteristics are chosen at provisioning time and can be adjusted later (with one exception — storage never shrinks).
Every cluster is provisioned with mandatory TLS and, optionally, a set of managed database users and Relay connections so your team can reach it through the governance layer immediately.