Skip to content
Klutch.sh Docs

Deploying a Langfuse App

Introduction

Langfuse is an open-source observability and analytics platform for LLM applications. Deploying Langfuse with a Dockerfile on Klutch.sh gives you reproducible builds, managed secrets, and persistent storage for telemetry and metadata—all configured from klutch.sh/app. This guide covers installation, repository prep, a production-ready Dockerfile, deployment steps, Nixpacks overrides, sample API usage, and production best practices.

Prerequisites

  • A Klutch.sh account (create one)
  • A GitHub repository containing your Langfuse code/config (GitHub is the only supported git source)
  • Docker familiarity and Node.js 18+ knowledge (Next.js-based)
  • PostgreSQL credentials
  • Storage for logs and optional cache

For onboarding, see the Quick Start.

Architecture and ports

  • Langfuse serves HTTP; set the internal container port to 3000.
  • PostgreSQL should run as a separate Klutch.sh TCP app, exposed on port 8000 and connected internally on 5432.
  • Persistent storage is optional (for logs/cache) but recommended if you store artifacts locally.

Repository layout

langfuse/
├── Dockerfile              # Must be at repo root for auto-detection
├── package.json
├── pnpm-lock.yaml          # or yarn.lock / package-lock.json
├── .env.example            # Template only; no secrets
└── README.md

Keep secrets out of Git; store them in Klutch.sh environment variables.

Installation (local) and starter commands

Install dependencies and run locally before pushing to GitHub:

Terminal window
pnpm install
pnpm build
pnpm start -- --port 3000

Optional helper start.sh for portability and Nixpacks fallback:

#!/usr/bin/env bash
set -euo pipefail
pnpm migrate:deploy || true
pnpm start -- --port 3000

Make it executable with chmod +x start.sh.

Dockerfile for Langfuse (production-ready)

Place this Dockerfile at the repo root; Klutch.sh auto-detects it (no Docker selection in the UI):

FROM node:18-alpine AS build
WORKDIR /app


COPY package.json pnpm-lock.yaml* yarn.lock* package-lock.json* ./
RUN corepack enable
RUN pnpm install --frozen-lockfile


COPY . .
RUN pnpm build


FROM node:18-alpine
WORKDIR /app
ENV NODE_ENV=production PORT=3000


COPY --from=build /app /app
RUN corepack enable && pnpm install --prod --frozen-lockfile


EXPOSE 3000
CMD ["pnpm", "start", "--", "--port", "3000"]

Notes:

  • Add build tools (apk add --no-cache python3 make g++) in the build stage if native modules are needed.
  • Keep Next.js output within the image; static assets are served directly.

Environment variables (Klutch.sh)

Set these in the Klutch.sh app settings (Secrets tab) before deploying:

  • NODE_ENV=production
  • PORT=3000
  • DATABASE_URL=postgres://<user>:<password>@<host>:<port>/<db>
  • NEXTAUTH_URL=https://example-app.klutch.sh
  • NEXTAUTH_SECRET=<secure-secret>
  • LANGFUSE_HOST=https://example-app.klutch.sh
  • LANGFUSE_PUBLIC_KEY=<public-key>
  • LANGFUSE_SECRET_KEY=<secret-key>

If you deploy without the Dockerfile and need Nixpacks overrides:

  • NIXPACKS_BUILD_CMD=pnpm install --frozen-lockfile && pnpm build
  • NIXPACKS_START_CMD=pnpm start -- --port 3000
  • NIXPACKS_NODE_VERSION=18

These keep Langfuse compatible with Nixpacks defaults when a Dockerfile is absent.

Attach persistent volumes

In Klutch.sh storage settings, add mount paths and sizes (no names required):

  • /app/.next/cache — optional cache for faster rebuilds.
  • /app/logs — optional if you store logs locally.

Ensure these paths are writable inside the container.

Deploy Langfuse on Klutch.sh (Dockerfile workflow)

  1. Push your repository (with the Dockerfile at the root) to GitHub.
  2. Open klutch.sh/app, create a project, and add an app.
  1. Connect the GitHub repository; Klutch.sh automatically detects the Dockerfile.
  2. Choose HTTP traffic for Langfuse.
  3. Set the internal port to 3000.
  4. Add the environment variables above (database, NextAuth, Langfuse keys, and any NIXPACKS_* overrides if you temporarily deploy without the Dockerfile).
  5. Attach persistent volumes for /app/.next/cache (and /app/logs if used), selecting sizes that fit your caching/logging needs.
  6. Deploy. Your Langfuse instance will be reachable at https://example-app.klutch.sh; attach a custom domain if desired.

Sample ingestion request

Send an event to Langfuse (replace placeholders):

Terminal window
curl -X POST "https://example-app.klutch.sh/api/public/ingest" \
  -H "Content-Type: application/json" \
  -H "x-langfuse-public-key: <public-key>" \
  -H "x-langfuse-secret-key: <secret-key>" \
  -d '{"events":[{"type":"trace","id":"trace_123","name":"demo-trace","timestamp":1690000000}]}'

Health checks and production tips

  • Add a reverse proxy probe to /api/health or a lightweight status route.
  • Enforce HTTPS at the edge; forward HTTP to port 3000 internally.
  • Keep lockfiles committed and Node version pinned; test upgrades before applying.
  • Monitor database performance and set connection pool limits to match your Postgres deployment.
  • Rotate keys and secrets regularly; store them only in Klutch.sh secrets.

Langfuse on Klutch.sh combines reproducible Docker builds with managed secrets, optional persistent cache/log storage, and flexible HTTP/TCP routing. With the Dockerfile at the repo root and port 3000 configured, you can run observability for your LLM apps without extra YAML or workflow overhead.