Analytics

Logs, metrics, traces and alerts, plus product analytics and ML workloads.

Get started

3 guides, in order
  1. Step 1Set Up Structured Logging for a Production AppEmit JSON logs from your Railway services so the Log Explorer can parse, color, and filter them by level and custom attributes.
  2. Step 2Ship Application Logs to a Third-Party AggregatorBuild a log drain on Railway with a Vector forwarder service. Apps send logs over private networking, and Vector fans them out to any aggregator.
  3. Step 3Self-Host Langfuse for LLM ObservabilityDeploy Langfuse v3 on Railway with Postgres, ClickHouse, Redis, and object storage. Trace, debug, and evaluate your LLM applications on infrastructure you control.

Build and deploy

7 guides
Deploy a Jupyter Server for Team Data WorkRun a shared JupyterLab server on Railway with token authentication, a persistent volume for notebooks, and health checks for zero-downtime deploys.Run ClickHouse for Product AnalyticsDeploy ClickHouse on Railway from the template or the official Docker image, size its volume for event data, and ingest and query analytics events from a Node.js service.Connect a Third-Party Observability ToolSend telemetry from your Railway services to third-party observability tools. Covers vendor SDKs (Datadog, New Relic, Sentry), OpenTelemetry (OTLP), and common deployment pitfalls.Set Up a Datadog Agent in RailwayLearn how to set up a Datadog agent in Railway.Deploy Triton Inference Server on RailwayLearn the benefits of using Triton inference server and how to deploy it on RailwayDeploy Offen on RailwayOffen is an open-source, lightweight, and free analytics solution. In this post, we’ll deploy an Offen instance connected to a PostgreSQL database on Railway!Self-hosted website analyticsIn this post, we talk about self-hosted alternatives to Google Analytics, why you should be using them instead and how you'd go about deploying them on Railway!

Run in production

3 guides
Self-Host Uptime Kuma for Uptime MonitoringDeploy Uptime Kuma on Railway with a volume for persistent monitor history. Monitor public sites and internal services, and publish status pages.Debug a Production Incident with Logs, Metrics, and TracesA step-by-step incident debugging workflow using Railway's built-in observability surfaces. Confirm scope with metrics, find the failure in HTTP and deployment logs, correlate requests, and mitigate with a rollback.Set Up Alerts for Crashes, Restarts, and Failed DeploysConfigure Railway webhooks and monitors to get notified in Slack, Discord, or your own service when a deployment fails, crashes, or a resource threshold is crossed.

More guides

5 guides
Instrument an App with OpenTelemetryEnable Railway's built-in tracing, follow requests from the edge into your service, and add OpenTelemetry spans from your code with automatic instrumentation or an SDK.Roll Back a Bad Deploy FastHow to revert a broken deployment on Railway in under a minute. Covers rollback vs redeploy vs restart, image retention limits, and healthcheck-gated deploys that stop bad code before it goes live.Right-Size CPU and Memory from Real MetricsHow to read the Metrics tab on Railway, translate real CPU and memory usage into a monthly cost, and set replica and usage limits that cap spend without crashing your service.Load Test an App Before Launch with k6Write a k6 load test, run it locally or as a Railway service, set realistic performance targets, and read Railway metrics while the test runs.Logs, Metrics, and Traces: What Does Each Signal Tell You?Learn how logs, metrics, and traces work together to cut incident resolution time. See how each signal answers a different question.