Instrument a TypeScript/JavaScript app with the Langfuse JS/TS SDK v5 (OpenTelemetry-based), including flushing in serverless and edge runtimes

Documented steps

1. The current GA major version is JS/TS SDK v5; v4 and older are Deprecated on Langfuse Cloud v4. Install the split packages: `npm install @langfuse/tracing @langfuse/otel @opentelemetry/sdk-node`. Add `@langfuse/client` for the prompts/datasets/scores APIs and `@langfuse/openai` to trace OpenAI calls.
2. Set env vars: `LANGFUSE_SECRET_KEY=sk-lf-...`, `LANGFUSE_PUBLIC_KEY=pk-lf-...`, `LANGFUSE_BASE_URL=https://cloud.langfuse.com` (regional: us.cloud.langfuse.com, jp.cloud.langfuse.com, hipaa.cloud.langfuse.com).
3. Create an `instrumentation.ts` entry point wiring OpenTelemetry to Langfuse: `import { NodeSDK } from "@opentelemetry/sdk-node"; import { LangfuseSpanProcessor } from "@langfuse/otel"; export const sdk = new NodeSDK({ spanProcessors: [new LangfuseSpanProcessor()] }); sdk.start();`. Import this file first at app startup, before any traced code runs.
4. Initialize a client for non-tracing APIs: `import { LangfuseClient } from "@langfuse/client"; const langfuse = new LangfuseClient();` (reads env vars), or pass `{ publicKey, secretKey, baseUrl }` explicitly.
5. Trace an OpenAI call with the dedicated wrapper: `import OpenAI from "openai"; import { observeOpenAI } from "@langfuse/openai"; const openai = observeOpenAI(new OpenAI()); const res = await openai.chat.completions.create({ messages: [...], model: "gpt-4o" });`. This auto-captures prompts, completions, latency, token usage, and errors.
6. For non-OpenAI or manual spans: `import { startActiveObservation } from "@langfuse/tracing"; await startActiveObservation("my-trace", async (span) => { span.update({ input: "...", output: "..." }); });`.
7. In serverless and edge runtimes (Vercel, AWS Lambda) the function can be frozen or recycled the instant your handler returns, so flush explicitly: `await main().finally(() => sdk.shutdown());`. The docs state shutdown flushes events and is required for short-lived applications.
8. Verified against current official docs on 2026-07-29: https://langfuse.com/docs/observability/sdk/typescript/overview , https://langfuse.com/integrations/model-providers/openai-js , https://langfuse.com/docs/compatibility