You can use the CLI to instrument your AWS Lambda functions with Datadog. The CLI enables instrumentation by modifying existing Lambda functions’ configuration and hence does not require redeployment. It is the quickest way to get started with Datadog serverless monitoring.

You can also add the command to your CI/CD pipelines to enable instrumentation for all your serverless applications. Run the command after your normal serverless application deployment, so that changes made by the Datadog CLI command do not get overridden.

Installation

To instrument your Lambda functions using the datadog-ci lambda instrument command, follow the instructions for a specific runtime listed below:

• .NET
• Go
• Java
• Node.js
• Python
• Ruby

Commands

instrument

Run datadog-ci lambda instrument to apply Datadog instrumentation to a Lambda. This command adds the Datadog Lambda Library and/or the Datadog Lambda Extension as Lambda Layers to the instrumented Lambda functions and modifies their configurations.

datadog-ci lambda instrument -f <function-name> -f <another-function-name> -r us-east-1 -v 81 -e 49

# Instrument multiple functions specified by full ARNs
datadog-ci lambda instrument -f <lambda-arn> -f <another-lambda-arn> -f <a-third-lambda-arn> -v 81 -e 49

# Instrument function(s) in interactive mode
datadog-ci lambda instrument -i

# Instrument multiple functions that match a regex pattern
datadog-ci lambda instrument --functions-regex <valid-regex-pattern> -r us-east-1 -v 81 -e 49

# Dry run of all updates
datadog-ci lambda instrument -f <function-name> -f <another-function-name> -r us-east-1 -v 81 -e 49 --dry-run

uninstrument

Run datadog-ci lambda uninstrument to revert Datadog instrumentation in a Lambda. This command automatically removes the Datadog configuration, such as the Datadog Lambda Library and the Datadog Lambda Extension layers, as well as other configurations applied by the datadog-ci.

# Uninstrument multiple functions specified by names
datadog-ci lambda uninstrument -f <function-name> -f <another-function-name> -r us-east-1

# Uninstrument function(s) in interactive mode
datadog-ci lambda uninstrument -i

# Uninstrument multiple functions that match a regex pattern
datadog-ci lambda uninstrument --functions-regex <valid-regex-pattern> -r us-east-1

# Dry run of all updates
datadog-ci lambda uninstrument -f <function-name> -f <another-function-name> -r us-east-1 --dry-run

See the configuration section for additional settings.

Configuration

AWS Credentials

You must have valid AWS credentials configured with access to the Lambda and CloudWatch services where you are running any datadog-ci lambda command.

Environment variables

You must expose these environment variables in the environment where you are running datadog-ci lambda instrument:

Arguments

Configuration can be done using command-line arguments or a JSON configuration file (see the next section).

instrument

You can pass the following arguments to instrument to specify its behavior. These arguments will override the values set in the configuration file, if any.

uninstrument

The following arguments are passed to uninstrument to specify its behavior. These arguments will override the values set in the configuration file, if any.

Any other argument stated on the instrument table, but not below, will be ignored, this to allow you to uninstrument quicker, if needed.

Configuration file

Instead of supplying arguments, you can create a configuration file in your project and simply run the datadog-ci lambda {instrument|uninstrument} --config datadog-ci.json command on each deployment. Specify the datadog-ci.json using the --config argument, and use this configuration file structure:

{
    "lambda": {
        "layerVersion": 10,
        "extensionVersion": 8,
        "functions": ["arn:aws:lambda:us-east-1:000000000000:function:autoinstrument"],
        "region": "us-east-1",
        "tracing": true,
        "mergeXrayTraces": true,
        "captureLambdaPayload": true,
        "forwarder": "arn:aws:lambda:us-east-1:000000000000:function:datadog-forwarder",
        "logLevel": "debug",
        "service": "some-service",
        "version": "b17s47h3w1n",
        "profile": "my-credentials",
        "environment": "staging",
        "extraTags": "layer:api,team:intake"
    }
}

Troubleshooting Lambda Instrumentation

To troubleshoot issues you encounter with Datadog monitoring on your Lambda functions, run the datadog-ci lambda flare command in the root of your project directory. This command collects important data about the Lambda function, such as environment variables and the config file. These files will be submitted to Datadog support via a ticket matching the provided Zendesk case ID.

Note: This command works whether or not your Lambda functions were instrumented using datadog-ci lambda instrument.

Examples

# Collect and send files to Datadog support for a single function
datadog-ci lambda flare -f <function-arn> -c <case-id> -e <email-on-case-id>

# Include recent CloudWatch logs
datadog-ci lambda flare -f <function-name> -r <AWS region> -c <case-id> -e <email-on-case-id> --with-logs

# Dry run: collect data, but don't send to Datadog support
datadog-ci lambda flare -f <function-arn> -c <case-id> -e <email-on-case-id> --with-logs --dry-run

Arguments

Permissions

To run this command, you must have the following IAM permissions:

• lambda:GetFunction
• lambda:ListTags
• logs:DescribeLogStreams (required if --with-logs is set)
• logs:GetLogEvents (required if --with-logs is set)

Community

For product feedback and questions, join the #serverless channel in the Datadog community on Slack.