AWS Lambda Debugging for Serverless Python

This page is deprecated. You can check the active version from here.
Using AWS Lambda Debugging with Python Lambda Functions
1.Integrate Thundra’s Python agent with your Lambda function using one of the available integration options.
2.Add the ptvsd layer is shown below, changing the ${region} parameter:arn:aws:lambda:${region}:269863060030:layer:thundra-lambda-python-ptvsd-layer:1
3.​Enable the AWS Lambda Debugging by adding the thundra_agent_lambda_debugger_auth_token environment variable and setting thundra_agent_lambda_debugger_enable to true.
4.​Specify a session name if needed (as long as you don't have multiple concurrent debugging sessions, you can use the default value and not specify a session name).
5.Configure Thundra Debugger in your IDE (VSCode and others).
6.Start the debug session in your IDE and invoke your function. Thundra Debugger will establish the connection between the IDE and your Lambda function using your authentication token and the session name.
7.You can disable the Online Debugging once you complete your debugging session.
For longer debugging sessions, you will need to make sure to increase your function's timeout value when you enable online debugging.
To use the online debugging feature, you need to use version 2.4.5 of Thundra’s Python agent along with layer version 20 or higher.
Enabling the AWS Lambda Debugging
The fastest way to enable the AWS Lambda Debugging for a Lambda function is to set the thundra_agent_lambda_debugger_auth_token environment variable to the authentication token that you receive from the Thundra console, and then to set thundra_agent_lambda_debugger_enable to true. By doing this, Thundra will start your function in debug mode and you will be able to connect your function from your IDE to start the debugging session.
Disabling the AWS Lambda Debugging
The AWS Lambda Debugging feature is disabled by default. However, if you want to manually disable it, you can set the thundra_agent_lambda_debugger_enable environment variable to false. This ensures that even if you have thundra_agent_lambda_debugger_auth_token, AWS Lambda Debugging will still be disabled.
Specifying a session name
In order to match the two ends of an AWS Lambda Debugging session (your Lambda function invocation and your local IDE), Thundra uses session names. When you enable AWS Lambda Debugging in your Lambda function, the session name is set to the predefined value "default." If you want to use another session name, you can specify it using the thundra_agent_lambda_debugger_session_name environment variable.
Broker Host Selection
Thundra’s broker establishes communication between your Lambda function and your IDE debugger by receiving data from the Lambda function and then forwarding it to your IDE. To reduce overhead during this transmission process, you need to select a broker host that provides the shortest pathway between your Lambda and IDE. Similarly, the broker host needs to be close to your Lambda function or your IDE location.
Thundra provides a broker host in Oregon by default, but we also support different broker hosts (listed below). You can select a broker host and set it to the thundra_agent_lambda_debugger_broker_host environment variable to reduce latency during a debugging session.
​debug.thundra.io (us-west-2 - Oregon)
​debug-us-east-1.thundra.io (us-east-1 - N. Virginia)
​debug-eu-west-2.thundra.io (eu-west-2 - London)
​debug-ap-northeast-1.thundra.io (ap-northeast-1 - Tokyo)
Concurrent Session Limit
The number of debugging sessions an account can perform is limited for now. Currently, you can start two concurrent debugging sessions at a time. You can contact us at [email protected] or via the intercom bubble on the right-hand bottom corner of the Thundra console if you want to increase your concurrent session limits.
Environment variables to configure AWS Lambda Debugging
Environment Variable
Description
thundra_agent_lambda_debugger_auth_token
The authentication token you receive from the Thundra console. This field is required and there is no default value.
thundra_agent_lambda_debugger_session_name
The unique session name that identifies a current debugging session from other concurrent debugging sessions. The default value is "default."
thundra_agent_lambda_debugger_broker_host
The broker host address to which Thundra connects to start a debugging session with your IDE. The default value is "debug.thundra.io."
thundra_agent_lambda_debugger_broker_port
The broker port that Thundra uses to connect the broker. The default value is 444.
thundra_agent_lambda_debugger_wait_max
The maximum amount of time in milliseconds that your function should wait until a debugging session request comes from the IDE. The default value is 60000.
thundra_agent_lambda_debugger_logs_enable
The Boolean value that enables Thundra debugger logs. The default value is  false.
thundra_agent_lambda_debugger_enable
The Boolean value that enables/disables the Thundra debugger. It can be set to true to enable the debugger (the authentication token environment variable should also be set), and it can be set to false to disable the Thundra debugger.

Python - Previous

Enrich Tracing in Python

Next - Python

FAQ for Python SDK

Last modified 3mo ago

Copy link

On this page

Using AWS Lambda Debugging with Python Lambda Functions

Enabling the AWS Lambda Debugging

Disabling the AWS Lambda Debugging

Specifying a session name

Broker Host Selection

Concurrent Session Limit