Arize has first-class support for LangChain applications. After instrumentation, you will have a full trace of every part of your LLM application, including input, embeddings, retrieval, functions, and output messages.
We follow a standardized format for how a trace data should be structured using openinference, which is our open source package based on OpenTelemetry. The package we are using is arize-otel, which is a lightweight convenience package to set up OpenTelemetry and send traces to Arize.
Use our code block below to get started using our LangChainInstrumentor.
# Import open-telemetry dependencies
from arize.otel import register
# Setup OTel via our convenience function
tracer_provider = register(
space_id = "your-space-id", # in app space settings page
api_key = "your-api-key", # in app space settings page
project_name = "your-project-name", # name this to whatever you would like
)
# Import the automatic instrumentor from OpenInference
from openinference.instrumentation.langchain import LangChainInstrumentor
# Finish automatic instrumentation
LangChainInstrumentor().instrument(tracer_provider=tracer_provider)
For more in-detail demonstration, check our Colab tutorial:
The example below utilizes the OpenInference JavaScript LangChain example.
Navigate to the backend folder.
In addition to the above package, sending traces to Arize requires the following package: @opentelemetry/exporter-trace-otlp-grpc. This package can be installed in your environment by running the following command in your shell.
instrumentation.ts should be implemented as below (you'll need to install all of the packages imported below in the same manner as above):
/*instrumentation.ts */
import { LangChainInstrumentation } from "@arizeai/openinference-instrumentation-langchain";
import { ConsoleSpanExporter } from "@opentelemetry/sdk-trace-base";
import {
NodeTracerProvider,
SimpleSpanProcessor,
} from "@opentelemetry/sdk-trace-node";
import { Resource } from "@opentelemetry/resources";
import { OTLPTraceExporter as GrpcOTLPTraceExporter } from "@opentelemetry/exporter-trace-otlp-grpc"; // Arize specific
import { diag, DiagConsoleLogger, DiagLogLevel } from "@opentelemetry/api";
import { Metadata } from "@grpc/grpc-js";
import * as CallbackManagerModule from "@langchain/core/callbacks/manager";
// For troubleshooting, set the log level to DiagLogLevel.DEBUG
diag.setLogger(new DiagConsoleLogger(), DiagLogLevel.DEBUG);
// Arize specific - Create metadata and add your headers
const metadata = new Metadata();
// Your Arize Space and API Keys, which can be found in the UI
metadata.set('space_id', 'your-space-id');
metadata.set('api_key', 'your-api-key');
const provider = new NodeTracerProvider({
resource: new Resource({
// Arize specific - The name of a new or preexisting model you
// want to export spans to
"model_id": "your-model-id",
"model_version": "your-model-version"
}),
});
provider.addSpanProcessor(new SimpleSpanProcessor(new ConsoleSpanExporter()));
provider.addSpanProcessor(
new SimpleSpanProcessor(
new GrpcOTLPTraceExporter({
url: "https://otlp.arize.com/v1",
metadata,
}),
),
);
const lcInstrumentation = new LangChainInstrumentation();
// LangChain must be manually instrumented as it doesn't have
// a traditional module structure
lcInstrumentation.manuallyInstrument(CallbackManagerModule);
provider.register();
If you simultaneously want to send spans to a Phoenix collector, you should also add the following code blocks, from the original instrumentation.ts files.
import {
OTLPTraceExporter as ProtoOTLPTraceExporter
} from "@opentelemetry/exporter-trace-otlp-proto";
// add as another SpanProcessor below the previous SpanProcessor
provider.addSpanProcessor(
new SimpleSpanProcessor(
new ProtoOTLPTraceExporter({
// This is the url where your phoenix server is running
url: "http://localhost:6006/v1/traces",
}),
),
);
Follow the steps from the backend and frontend readme. Or simply run:
docker compose up --build
to build run the frontend, backend, and Phoenix all at the same time. Navigate to localhost:3000 to begin sending messages to the chatbot and check out your traces in Arize at app.arize.com or Phoenix at localhost:6006.
Native Thread Tracking
Arize supports native thread tracking with LangChain by enabling the use of session_id, thread_id, or conversation_id to group related calls. This flexibility allows for seamless tracking of multi-turn conversations, making it easier to monitor and analyze chatbot interactions for performance or debug issues using Arize's observability tools. Below is an example demonstrating how to set a thread_id in the metadata for a chat:
import { ChatOpenAI } from "@langchain/openai";
const chatModel = new ChatOpenAI({
openAIApiKey: "my-api-key",
modelName: "gpt-3.5-turbo",
});
async function run() {
// First message invocation
const response1 = await chatModel.invoke("Hello, how are you?", {
metadata: {
thread_id: "thread-456",
},
});
// Second message invocation
const response2 = await chatModel.invoke("What can you do?", {
metadata: {
thread_id: "thread-456",
},
});
// Print thread_id and session_id from the responses
console.log("Response 1 Metadata:");
console.log(`Thread ID: thread-456`);
console.log(`Session ID: ${response1.metadata?.session_id || "Not available"}`);
console.log("Response 2 Metadata:");
console.log(`Thread ID: thread-456`);
console.log(`Session ID: ${response2.metadata?.session_id || "Not available"}`);
}
run().catch(console.error);