Deploy OpenLIT observability
Protected OpenLIT AI observability with persistent ClickHouse.
Application
OpenLIT Gateway
Just deployed
OpenLIT
Just deployed
/app/client/data
OpenLIT ClickHouse
Just deployed
/var/lib/clickhouse
Deploy and Host OpenLIT on Railway
OpenLIT captures and explores traces, metrics, and logs from AI applications. This template adds persistent ClickHouse storage and one generated-credential gateway for the dashboard and OTLP/HTTP ingestion.
About Hosting OpenLIT
Railway exposes only Caddy over public HTTPS. OpenLIT, its embedded collector, ClickHouse, OTLP gRPC, and OpAMP stay on the private network, while separate generated credentials protect browser access and HTTP telemetry ingestion.
Why Deploy OpenLIT on Railway
The deployment provisions the application, two persistent volumes, private networking, TLS, generated secrets, and the upstream collector pipeline together. A protocol-level smoke can then send a trace and read it through OpenLIT's authenticated API, which proves the full SDK-to-dashboard path.
Common Use Cases
- Inspect LLM and agent traces, latency, token use, and cost during development.
- Give a small internal team one persistent AI observability workspace.
- Receive standard OTLP/HTTP telemetry from OpenLIT or OpenTelemetry SDKs.
Dependencies for OpenLIT Hosting
Deployment Dependencies
- OpenLIT
1.24.2 - ClickHouse
24.4.1 - Caddy
2.10 - Two 5 GB Railway volumes; expand ClickHouse first as telemetry grows
Template Content
OpenLIT Gateway
caddy:2.10-alpineOpenLIT ClickHouse
clickhouse/clickhouse-server:24.4.1