AWS Lambda - Python Function Instrumentation

This document covers how to install and configure OpenTelemetry distributed tracing for AWS Lambda functions written in Python and send data to Sumo Logic. To obtain telemetry data from AWS Lambda functions developed in Python language, you can use the Sumo Logic Distribution for OpenTelemetry Python Lambda, which provides auto-instrumentation.

Sumo Logic OTel Python Lambda layer supports:

• Python version between 3.7 and 3.11
• x86_64 and arm64 architectures

Sumo Logic Distribution for OpenTelemetry Lambda Layer

Sumo Logic Distribution for OpenTelemetry Lambda Layer version 1.20.0 provides packed OpenTelemetry Python libraries that automatically instrument Lambda functions. The biggest advantage of installing Sumo Logic OTel Lambda as a layer is disabling/enabling instrumentation of the Lambda function without changing the code.

Lambda function requirements

It is very simple to instrument your AWS Python Lambda function using the Sumo Logic Distro for Lambda layer. You'll need the following:

• Python version between 3.7 and 3.10
• Lambda layers add permissions
• Sumo Logic OTLP/HTTP Source endpoint URL - To send spans from the instrumented Lambda function to Sumo Logic you need an endpoint URL from an existing or new OTLP/HTTP source.

Deployment

1. Navigate to functions in the AWS Lambda Console and open the function you want to instrument.

2. Navigate to the Layers section and click Add a layer.

3. In the Choose a layer menu, select Specify an ARN and paste the ARN ID for your Lambda function AWS Region. Reference the table in the section Sumo Logic AWS Distro Lambda layers for AWS Region - amd64 (x86_64) architecture for the ARN ID.
   

4. Ensure the AWS Distro layer is present in the Layers section:
   

   note

   You must place the Sumo Logic layer first in the order or you won't see traces in Sumo Logic.

5. Navigate to the Configuration > Environment variables section and set up the following environment variables:

   • AWS_LAMBDA_EXEC_WRAPPER= /opt/otel-instrument - Enables auto-instrumentation.
   • OTEL_TRACES_SAMPLER = always_on - Enables traces sampling.
   • OTEL_SERVICE_NAME = YOUR_SERVICE_NAME - Sets the tracing service name in Sumo Logic. Make sure to define it as a string value that represents the function name and its business logic such as "Check SQS Lambda".
   • OTEL_RESOURCE_ATTRIBUTES - Sets OpenTelemetry resources. Add the deployment.environment=[environment-name] tag as needed to allow for filtering by environment on dashboard panels. (For more information, see Services Dashboard Panels). Tracing application and cloud.account.id are set with the OTEL_RESOURCE_ATTRIBUTES environment variable:
     • application=YOUR_APPLICATION_NAME - The string value, if the function is a part of complex system/application then set it for all other functions/applications.
     • cloud.account.id=YOUR_CLOUD_ACCOUNT_ID - Set an additional tag that will contain your AWS Lambda Account ID. This will help to provide more relevant data. All of the attributes above are comma separated key/value pairs (this is also a way to add additional information to the spans, just after comma add additional key=value pair) such as: OTEL_RESOURCE_ATTRIBUTES=application=YOUR_APPLICATION_NAME,cloud.account.id=123456789012.
   • SUMO_OTLP_HTTP_ENDPOINT_URL has to be set to send all gathered telemetry data to Sumo Logic. The URL comes from an OTLP/HTTP source. You can use an existing Source or create a new one, if needed.
   note

   The SUMOLOGIC_HTTP_TRACES_ENDPOINT_URL environment variable is deprecated. You'll need to switch from the HTTP Traces Source to OTLP/HTTP source and use the SUMO_OTLP_HTTP_ENDPOINT_URL environment variable instead.

   

6. Make sure you have X-Ray Tracing disabled in Lambda API Stage. Navigate to AWS API Gateway console, find your API and go to Stages. In the Logs/Tracing tab uncheck Enable X-Ray Tracing option.

   note

   If for whatever reason you cannot disable this, configure X-Ray context propagation by setting OTEL_PROPAGATORS=xray environment variable on your client side.

7. Your function should be successfully instrumented. Invoke the function and find your traces in the Sumo Logic Tracing screen.

Sumo Logic Distro Lambda layers for AWS Region - amd64 (x86_64) architecture

Go back to Step 3 (In the Choose a layer menu...).

Region	ARN
af-south-1	arn:aws:lambda:af-south-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
ap-east-1	arn:aws:lambda:ap-east-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
ap-northeast-1	arn:aws:lambda:ap-northeast-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
ap-northeast-2	arn:aws:lambda:ap-northeast-2:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
ap-northeast-3	arn:aws:lambda:ap-northeast-3:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
ap-south-1	arn:aws:lambda:ap-south-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
ap-southeast-1	arn:aws:lambda:ap-southeast-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
ap-southeast-2	arn:aws:lambda:ap-southeast-2:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
ca-central-1	arn:aws:lambda:ca-central-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
eu-central-1	arn:aws:lambda:eu-central-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
eu-north-1	arn:aws:lambda:eu-north-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
eu-south-1	arn:aws:lambda:eu-south-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
eu-west-1	arn:aws:lambda:eu-west-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
eu-west-2	arn:aws:lambda:eu-west-2:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
eu-west-3	arn:aws:lambda:eu-west-3:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
me-south-1	arn:aws:lambda:me-south-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
sa-east-1	arn:aws:lambda:sa-east-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
us-east-1	arn:aws:lambda:us-east-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
us-east-2	arn:aws:lambda:us-east-2:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
us-west-1	arn:aws:lambda:us-west-1:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1
us-west-2	arn:aws:lambda:us-west-2:663229565520:layer:sumologic-otel-lambda-python-x86_64-v1-20-0:1

Sumo Logic Distro Lambda layers for AWS Region - arm64 (arm) architecture

Go back to Step 3 (In the Choose a layer menu...).

Region	ARN
ap-northeast-1	arn:aws:lambda:ap-northeast-1:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
ap-northeast-3	arn:aws:lambda:ap-northeast-3:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
ap-south-1	arn:aws:lambda:ap-south-1:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
ap-southeast-1	arn:aws:lambda:ap-southeast-1:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
ap-southeast-2	arn:aws:lambda:ap-southeast-2:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
eu-central-1	arn:aws:lambda:eu-central-1:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
eu-west-1	arn:aws:lambda:eu-west-1:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
eu-west-2	arn:aws:lambda:eu-west-2:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
us-east-1	arn:aws:lambda:us-east-1:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
us-east-2	arn:aws:lambda:us-east-2:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1
us-west-2	arn:aws:lambda:us-west-2:663229565520:layer:sumologic-otel-lambda-python-arm64-v1-20-0:1

Sumo Logic OTel Lambda container instrumentation

Sumo Logic Distribution for OpenTelemetry Lambda Layer version 1.20.0 also provides packed OpenTelemetry Python libraries for container based Lambda functions.

note

The instructions below support only AWS Base Images for Lambda.

Container based lambda requirements

Instrumentation of container based AWS Lambda function requires some changes in the Dockerfile and image rebuild. You'll need the following:

• Docker
• Python version between 3.7 and 3.11
• Sumo Logic OTLP/HTTP Source endpoint URL - To send spans from the instrumented Lambda function to Sumo Logic you need an endpoint URL from an existing or new OTLP/HTTP source.

Lambda function image changes

1. Download and extract Sumo Logic Distribution for OpenTelemetry Java Lambda Layer archive with instrumentation packages specific for your architecture - amd64 (x86_64) or arm64.

2. Extracted instrumentation libraries have to be added to the image in /opt directory. Please see Dockerfile example:

   FROM public.ecr.aws/lambda/python:3.8-arm64
   
   # Lambda Function Code  
   COPY lambda_function.py ${LAMBDA_TASK_ROOT}  
   COPY requirements.txt  .  
   RUN  pip3 install -r requirements.txt --target "${LAMBDA_TASK_ROOT}"
   
   **# Copy OT Instrumentation  
   COPY collector-config/ /opt/collector-config/  
   COPY extensions/ /opt/extensions/  
   COPY python/ /opt/python/  
   COPY otel-instrument /opt/**
   
   CMD \[ "lambda_function.lambda_handler" \]

3. Rebuild docker image.

Deployment

1. Navigate to functions in the AWS Lambda Console and open the function you want to instrument.

2. Deploy new function image.

3. Navigate to the Configuration > Environment variables section and set up the following environment variables:

   • AWS_LAMBDA_EXEC_WRAPPER = /opt/otel-instrument - Enables auto-instrumentation.
   • OTEL_TRACES_SAMPLER = always_on - enables traces sampling.
   • OTEL_SERVICE_NAME = YOUR_SERVICE_NAME - Ensure you define it as a string value that represents the function name and its business logic such as "Check SQS Lambda". This will appear as the tracing service name in Sumo Logic.
   • OTEL_RESOURCE_ATTRIBUTES - Sets OpenTelemetry resources. Add the deployment.environment=[environment-name] tag as needed to allow for filtering by environment on dashboard panels. (For more information, see Services Dashboard Panels). Tracing application and cloud.account.id are set with the OTEL_RESOURCE_ATTRIBUTES environment variable:
     • application=YOUR_APPLICATION_NAME - the string value, if the function is a part of complex system/application then set it for all other functions/applications.
     • cloud.account.id=YOUR_CLOUD_ACCOUNT_ID - set an additional tag that will contain your AWS Lambda Account ID. This will help to provide more relevant data. All of the attributes above are comma separated key/value pairs (this is also a way to add additional information to the spans, just after comma add additional key=value pair), such as OTEL_RESOURCE_ATTRIBUTES=application=YOUR_APPLICATION_NAME,cloud.account.id=123456789012.
   • SUMO_OTLP_HTTP_ENDPOINT_URL has to be set to send all gathered telemetry data to Sumo Logic. The URL comes from an OTLP/HTTP source. You can use an existing Source or create a new one, if needed.
   note

   The SUMOLOGIC_HTTP_TRACES_ENDPOINT_URL environment variable is deprecated. You'll need to switch from the HTTP Traces Source to OTLP/HTTP source and use the SUMO_OTLP_HTTP_ENDPOINT_URL environment variable instead.

   

4. Your function should be successfully instrumented. Invoke the function and find your traces in the Sumo Logic Tracing screen.

Context propagation

In case of an external request to the Lambda function, it is important to propagate the context. Enabling AWS X-Ray context propagation on the client side will help to visualize the complex flow of the trace.

For applications instrumented by OpenTelemetry SDK, it is enough to install AWS X-Ray propagator dependency specific for an instrumentation and configure the OTEL_PROPAGATORS environment variable (for example: export OTEL_PROPAGATORS= tracecontext,baggage,xray).