Skip to main content

​
AWS Lambda

For AWS Lambda, Datadog provides a dedicated library, datadog-lambda-js, which wraps dd-trace and adds Lambda-specific functionality:
  • Cold start detection and tagging
  • Automatic flushing of traces before the Lambda function returns
  • Lambda layer support for zero-code-change installation
  • Enhanced metrics via the Datadog Lambda Extension
The easiest approach is to add the Datadog Lambda layer to your function. The layer includes both the Datadog Agent (as a Lambda Extension) and dd-trace pre-configured for Lambda. For manual layer installation, see the Datadog Lambda documentation.

​
Install via npm

npm
npm install datadog-lambda-js dd-trace
yarn
yarn add datadog-lambda-js dd-trace

​
Usage

Wrap your Lambda handler with datadog-lambda-js:
handler.js
'use strict'

const { datadog } = require('datadog-lambda-js')
const tracer = require('dd-trace').init()

// Wrap your handler with datadog() to enable:
// - automatic trace/log correlation
// - enhanced metrics
// - automatic trace flushing before the function returns
const myHandler = async (event, context) => {
  return {
    statusCode: 200,
    body: JSON.stringify({ message: 'Hello from Lambda!' }),
  }
}

module.exports.handler = datadog(myHandler)

​
Key Lambda considerations

Without the Datadog Lambda Extension or a Datadog Agent in the same environment, traces cannot be forwarded to Datadog. The datadog-lambda-js wrapper handles flushing traces synchronously before your Lambda function’s execution context is frozen.
Flush interval: In Lambda, dd-trace automatically sets flushInterval to 0 when AWS_LAMBDA_FUNCTION_NAME is set, ensuring traces are flushed synchronously at the end of each invocation rather than on a timer. Cold starts: The tracer adds some initialization overhead. Lambda layers and the Datadog Extension minimize cold start impact. Agent connectivity: The Datadog Lambda Extension runs in the same execution environment as your function and listens on localhost:8126, so no network configuration is needed when using the extension.

​
Environment variables for Lambda

DD_SERVICE=my-lambda-function
DD_ENV=production
DD_VERSION=1.0.0
DD_API_KEY=<your-datadog-api-key>
DD_SITE=datadoghq.com
When using the Datadog Extension, set DD_API_KEY (or use AWS Secrets Manager) and DD_SITE to route data to the correct Datadog region.

​
Google Cloud Functions

dd-trace automatically detects Google Cloud Functions via the FUNCTION_NAME (deprecated runtimes) or K_SERVICE (newer runtimes) environment variables. When detected:
  • profiling.enabled defaults to false
  • telemetry.enabled defaults to false
  • remoteConfig.enabled defaults to false
  • The service name defaults to the function name
Install and initialize the same way as a regular Node.js application:
index.js
require('dd-trace').init()

exports.helloWorld = (req, res) => {
  res.send('Hello, World!')
}

​
Azure Functions

dd-trace automatically detects Azure Functions via the WEBSITE_SITE_NAME environment variable. When detected, serverless defaults are applied (same as Lambda/GCF). dd-trace includes a built-in azure-functions plugin:
index.js
require('dd-trace').init()

module.exports = async function (context, req) {
  context.res = {
    body: 'Hello from Azure Functions!',
  }
}
The azure-functions plugin is listed in the available integrations and is enabled by default.

​
Serverless defaults

When dd-trace detects a serverless environment (Lambda, GCF, or Azure Functions), it applies these defaults automatically:
OptionDefault in serverlessNotes
crashtracking.enabledfalseCrash tracking requires a long-lived process
profiling.enabledfalseProfiling overhead is not suitable for short-lived functions
telemetry.enabledfalseReduces noise from short-lived invocations
remoteConfig.enabledfalseRemote config requires a persistent connection
flushInterval0 (Lambda only)Flush synchronously before execution context freezes
Service nameFunction/service nameInferred from AWS_LAMBDA_FUNCTION_NAME, K_SERVICE, or WEBSITE_SITE_NAME

​
Issues specific to Lambda

For issues specific to Lambda integration (layer configuration, extension problems, metric forwarding), open an issue in the datadog-lambda-js repository. For core instrumentation issues not related to the Lambda runtime, open an issue in the dd-trace-js repository.

Build docs developers (and LLMs) love

Get started for freeTalk to us