Skip to main content
OpenLIT is lightweight to self-host. It requires three components:
  1. OpenLIT — the platform UI and backend
  2. ClickHouse — for storing telemetry data
  3. OpenTelemetry Collector — for receiving traces and metrics
If you already have ClickHouse or an OpenTelemetry Collector, you can reuse them. OpenLIT can connect to your existing infrastructure.

Docker Compose

Quickest way to get started — deploys all three components with one command

Kubernetes

Production-ready deployment using the OpenLIT Helm chart

Docker Compose

Docker Compose is the fastest way to run OpenLIT locally or on a single server.
1

Clone the repository

2

Start the stack

From the root directory of the repository:
This starts OpenLIT (port 3000), ClickHouse (port 8123), and the OpenTelemetry Collector (ports 4317 and 4318).
3

Open the UI

Navigate to http://127.0.0.1:3000 and log in with the default credentials:
Change the default credentials immediately for any non-local deployment.

Upgrading Docker Compose

When upgrading from versions prior to 1.15.0 (which introduced Fleet Hub), use the --remove-orphans flag:
This removes the now-obsolete standalone otel-collector container, which conflicts with the new integrated collector on ports 4317 and 4318.

Kubernetes

Deploy OpenLIT on Kubernetes using the official Helm chart. The chart deploys all three components.
1

Add the Helm repository

2

Install the Helm chart

To install in a specific namespace:
3

Verify the deployment

Ports

Next steps

Quickstart

Get LLM observability running in under 5 minutes

Configuration

Configure environment variables, database connections, and OAuth

Dashboards

Create custom visualizations for your AI telemetry data

Kubernetes Operator

Auto-inject instrumentation into workloads without code changes