Langfuse Otel Job stalls

[aesher9o1]

I started pushing data to langfuse via otel and i am noticing my jobs have stalled. The integration works okay with langfuse SDK and OTEL at low throughput though

[jannikmaierhoefer]

Hi @aesher9o1, could you share more information on your setup? cc @inkeep

[inkeep-ai-assistant[bot]]

To help diagnose the stalling issue when pushing data to Langfuse via OpenTelemetry at higher throughput, here's what would be helpful to share about your setup:

Environment Configuration:

• Which OTEL endpoint are you using? The endpoint should be set to LANGFUSE_BASE_URL + "/api/public/otel" with proper authentication headers(1)(2)
• Are you running in a serverless environment? If so, ensure you're calling flush operations after requests complete(3)
• What's your current throughput/volume when the stalling occurs?

OpenTelemetry Setup:

• Which SDK are you using (Python, JS/TS, or direct OTEL SDK)?
• How is your TracerProvider configured?(1)
• Are you using SimpleSpanProcessor or BatchSpanProcessor?(1)

Common Issues to Check:

1. Connection Configuration - Verify your OTEL exporter endpoint is correctly configured. The error logs in (2) show connection refused errors when OTEL_EXPORTER_OTLP_ENDPOINT isn't properly set to the Langfuse endpoint

2. Span Processor Flushing - In serverless or high-throughput scenarios, ensure proper flushing. For JS/TS, you should schedule flush after requests finish(3):

after(async () => await langfuseSpanProcessor.forceFlush());

(3)

3. Missing Dependencies - Ensure opentelemetry-exporter-otlp is installed in your runtime environment(4)

Could you share which SDK/language you're using and your current OpenTelemetry configuration? This will help identify the specific cause of the stalling.

📚 Sources:

1. Example: Using OpenTelemetry SDK with Langfuse OTel API
2. bug: Noticing errors in Langfuse sdk logs
3. TypeScript SDK - Instrumentation
4. bug: No module named 'opentelemetry.exporter.otlp'

---

Have another question? Just tag @inkeep.