Deployment

These pages take Andromeda from a first run to real use.

Start with Choosing a setup, which describes the parts every deployment has and the ones you can add. The Deployment planner asks about yours, in a short Simple run or an Advanced one that reaches every setting, and gives back the steps for your case, the warnings that apply and the files to run the broker with. Its Check my configuration mode reads a configuration you already have and reports what the broker would refuse or ignore.

One broker holding all the data is the simplest arrangement, but not the only one. Topologies covers the others, up to whole domains federated behind one broker, each with a Docker Compose example.

The broker keeps its data in one PostgreSQL server with PostGIS and TimescaleDB. PostgreSQL lists what the broker needs from that server, and PostgreSQL image describes the image published to provide it. With many tenants, PgBouncer lets one broker keep hundreds of them loaded.

Andromeda doesn't authenticate callers, so Security shows how to put a reverse proxy or API gateway in front and keep the broker on a network only the proxy reaches. TLS covers terminating TLS in the broker or at the proxy. HTTP versions covers HTTP/1.1, HTTP/2 and the experimental HTTP/3, and Browser clients what to change when a web application talks to the broker.

Andromeda v1.0 runs one broker process per database. Running several on the same database is still a work in progress, and Replicas describes where it stands.

Troubleshooting covers the problems a running deployment meets, most of them to do with database connections.

What the broker does once it runs is under Operations, and Configuration lists every setting.