← Templates · Monitoring & Analytics
AP
Arize Phoenix
CommunityOpen-source LLM observability — tracing, evals, and prompt/response inspection for AI apps, over OpenTelemetry.
| Runs | arizephoenix/phoenix:version-20.23.0 |
|---|---|
| Stack | python |
| Provisioned for it | Nothing, it's self-contained |
| Verified | Not yet |
Why deploy on Railyard
Railyard runs it standalone on a server you already pay for — no addons to provision, with zero-downtime deploys and scaling still included.
Deploying Arize Phoenix is one piece of a bigger stack you can keep building — add a database, a worker, a cache, or another app, all on the same servers you already pay for.
Good for
- LLM observability — tracing, evals, and prompt/response inspection for LLM-backed apps
- Debugging RAG pipelines by inspecting retrieved context against generated answers
- A self-hosted, OpenTelemetry-based alternative for LLM tracing
Good to know
- Ships with SQLite storage on a single volume — fine for a team's own traces; set PHOENIX_SQL_DATABASE_URL to a Postgres URL after deploy if you need concurrent writers
- No authentication is enabled by default — don't leave this publicly reachable without putting it behind your own access control
- The OTLP/gRPC collector endpoint (4317) isn't this app's public port — send traces over OTLP/HTTP to this app's URL at /v1/traces instead