Deployment and operations

High availability

Run several stateless replicas against one PostgreSQL database, so the gateway is never a single point of failure.

What it is

Run several stateless replicas behind a load balancer, sharing one PostgreSQL database. Sessions, pending sign-ins, quotas and pool health work across replicas.

Why you want it

An AI gateway in the request path of every tool cannot be a single point of failure.

How it works

  • Replicas heartbeat into the database, and the admin console shows live nodes against the licensed count.
  • Sessions, pending sign-ins, quotas, rate limits and pool health all work across replicas.
  • Configuration changes reach the other replicas within five seconds by default.
  • Licenses are counted per node: Community covers one node, and Business covers three per gateway.

Related features

A single binary

The gateway and its admin console ship as one signed binary, container image and Helm chart.

Details →

Linux server installer

One command installs Janus as a service with HTTPS, so one VM can run a production gateway.

Details →

Docker, Compose and Helm

Run the container, the Compose stack with Prometheus and Grafana, or the Helm chart.

Details →

See it on your own network.

The Community edition is free for up to 25 people. The 30-day Business trial unlocks every Business feature.

Download free Start a Business trial