Grafana 13.1.7 with PostgreSQL 16 as its own database deploys in one click onto a dedicated server Selfhost.dev provisions, secures and runs. Grafana is pointed at Postgres before it first starts, so every dashboard, user and alert rule is stored there from the beginning. The database is never reachable from the internet.
Two containers on one server, nothing to wire by hand. Six locations on three continents.
Grafana OSS keeping its own data in PostgreSQL instead of the SQLite file it ships with. It is not the PostgreSQL data source, which every Grafana has built in, this one included.
Teams putting Grafana in front of more than one person who want its dashboards, users and alert rules in a real database. Grafana's own docs recommend MySQL or PostgreSQL for production.
When the instance matters enough to back up and restore like any other database. For a trial or one person's dashboards, plain Grafana is one container less.
Many Grafana PostgreSQL tutorials end in a Docker Compose file like this: Postgres beside Grafana with port 5432 published. Some never set GF_DATABASE_TYPE, so Grafana still keeps its dashboards in SQLite. This template wires Grafana to its Postgres before the first boot and keeps the database off the internet.
Grafana uses SQLite unless GF_DATABASE_TYPE says otherwise, even with Postgres running next to it. Postgres is then only a data source while dashboards sit in a file on the Grafana volume. Here Grafana's database settings point at the bundled Postgres from the start.
Your team and your alert receivers
TLS and routing
Grafana 13.1.7
PostgreSQL 16
grafana-data
Admin password
| Grafana and Postgres on Docker Compose, by hand | Grafana PostgreSQL on Selfhost.dev |
|---|---|
Run grafana-enterprise:latest and postgres:16 | grafana/grafana:13.1.7, pinned, beside postgres:16-alpine |
Add GF_DATABASE_* or Grafana stays on SQLite | Set before first boot: Postgres holds dashboards, users and alert state |
Publish 5432:5432 to reach the database | Port 5432 stays inside the server; the firewall opens only 22, 80 and 443 |
Sign in with admin/admin, then change it | GF_SECURITY_ADMIN_USER and GF_SECURITY_ADMIN_PASSWORD set before first boot |
| Mount two volumes and keep them | postgresql-data for the database, grafana-data for plugins and renders |
| Put a proxy and a certificate in front | TLS per hostname, renewed for you; GF_SERVER_ROOT_URL follows your domain |
Grafana PostgreSQL requirements are modest: two light containers and two volumes behind HTTPS. Grafana's own database holds dashboards, users and alert state, not your metrics, so it stays small. 4 GB, the smallest server we offer, is the size we recommend.
postgres:16-alpine) in its own containerTwo containers, one server, no SQL to run.
The Launch button opens this exact template in the console. Pick one of six locations and submit; the template deploys in about five minutes.
The console shows the Grafana username and password, both set before first boot. Store them somewhere safe; the database needs no setup from you.
Sign in, add a data source and build or import a dashboard. Everything you create is written to the bundled Postgres.
Grafana with PostgreSQL costs the same to run as Grafana alone: both containers share one server. A European 4 GB server runs them at $0.02/hr, about $15 to $16 a month, billed by the hour against prepaid credits. No database fee, no fee per user.
| Item | Cost |
|---|---|
| Grafana OSS and PostgreSQL 16 | $0 |
| Users, editors and viewers | $0 at any number |
| Bandwidth | $0, traffic is included in the server price |
| Optional daily backups | +20% of the server price |
| Server | RAM, cores, disk | Where | Rate | About a month |
|---|---|---|---|---|
cx23 (recommended) | 4 GB, 2 cores, 40 GB | Europe | $0.02/hr | $15.18 |
cax11 (ARM) | 4 GB, 2 cores, 40 GB | Europe | $0.02/hr | $16.35 |
cx33 (room to grow) | 8 GB, 4 cores, 80 GB | Europe | $0.03/hr | $23.36 |
cpx21 | 4 GB, 3 cores, 80 GB | US | $0.12/hr | $87.75 |
cpx22 | 4 GB, 2 cores, 80 GB | Singapore | $0.10/hr | $72.56 |
Monthly figures assume 730 hours.
With a card on file or a past top-up, the server is powered off and held for 24 hours. We email you; a top-up in that window starts it again. Welcome-credit-only servers are deleted at zero, database included, so keep your own pg_dump.
Compare sizes and locations in the calculator, no signup. Per-user Grafana Cloud and Amazon prices sit on the Grafana hosting page.
Grafana PostgreSQL or SQLite is the one storage choice Grafana leaves to you. SQLite ships inside the binary and, in Grafana's own words, suits local development and small evaluation instances. For production its docs recommend MySQL or PostgreSQL. Only where the state lives changes.
| Where they differ | Grafana with SQLite | Grafana with PostgreSQL (this page) |
|---|---|---|
| Where its state lives | A file on the Grafana volume | Its own PostgreSQL 16 container |
| Grafana's guidance | Local development, small evaluation | Production |
| Your own copy | Dashboard JSON exports | pg_dump of the whole database |
| High availability | Not possible: HA needs a shared database | The kind of database HA needs; one Grafana server here |
Guidance quoted from Grafana's installation and high availability docs.
For a trial or one person's dashboards, Grafana with SQLite does the same job with one container less. Start there; dashboards move across as JSON when the instance grows into a team tool.
admin/admin. The administrator is seeded before the first start.The Grafana (Postgres) template deploys in about five minutes. That is the platform's own budget, not an estimate.
Nothing here in your way? The template deploys in about five minutes.
Launch Grafana (Postgres)Grafana PostgreSQL backup works two ways here. The server can take an optional daily snapshot of everything on it. With container access over SSH, which we grant on request, pg_dump inside the database container writes Grafana's whole state to one file you can keep anywhere.
| Method | What it covers | Where it lives |
|---|---|---|
| Daily server backup (optional, +20%) | Grafana, Postgres, plugins and renders; the last 7 kept | With the server; deleted with it |
pg_dump over SSH (container access on request) | Dashboards, folders, users, alert rules and alert state | Wherever you copy it |
Tell us how many dashboards, viewers and alert rules you run. We reply with the size to run and what it costs a month.
No, not by itself. Grafana's high availability design runs two or more Grafana servers behind a load balancer, all sharing one MySQL or PostgreSQL database. This template runs one Grafana server with its state already in PostgreSQL, the kind of database that design needs. The database has no public port, so a second Grafana server cannot reach it as deployed.
Grafana's documentation says SQLite isn't recommended for production and points to MySQL or PostgreSQL instead. SQLite works well for local development and small evaluation instances. Once dashboards and alert rules matter to a whole team, PostgreSQL is the safer home for them. This template sets it up before Grafana first starts.
No. This template uses PostgreSQL as Grafana's own database, where it keeps dashboards, users and alert state. The PostgreSQL data source is separate: it is built into every Grafana, this one included. It charts data from any Postgres database Grafana can reach. Use both if your metrics already live in Postgres.
Export each dashboard as JSON, deploy this template, add your data sources again, then import the files. Recreate users, teams and alert rules in the new instance. Your SQLite deployment keeps running until you delete it, so you can compare the two side by side before pointing your domain at the new one.
Grafana supports PostgreSQL 12 and newer for its own database, following PostgreSQL's upstream support timeline. This template runs postgres:16-alpine, inside that range, next to grafana/grafana:13.1.7. Grafana is pinned to that exact tag and Postgres stays on major version 16, so a restart moves neither.
Selfhost.dev is a managed self hosting platform for developers and small teams. Open source apps and databases on dedicated servers you control, billed by the hour, in six locations across three continents.
Best for: Grafana, Metabase, Elasticsearch with Kibana and managed PostgreSQL, MySQL, Redis, ClickHouse, OpenSearch and Kafka.