Integrate Thundra’s Python agent with your Lambda function using one of the available integration options.
ptvsd layer shown below, changing the
Enable the online debugging by adding the
thundra_agent_lambda_debugger_auth_token environment variable and setting
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).
Configure Thundra Debugger in your IDE (VSCode and others).
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.
You can disable the Online Debugging once you complete your debugging session.
The fastest way to enable the Thundra Debugger 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
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.
The online 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, online debugging will still be disabled.
In order to match the two ends of an online debugging session (your Lambda function invocation and your local IDE), Thundra uses session names. When you enable online 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.
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)
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.
The authentication token you receive from the Thundra console. This field is required and there is no default value.
The unique session name that identifies a current debugging session from other concurrent debugging sessions. The default value is "default."
The broker host address to which Thundra connects to start a debugging session with your IDE. The default value is "debug.thundra.io."
The broker port that Thundra uses to connect the broker. The default value is 444.
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.
The Boolean value that enables Thundra debugger logs. The default value is false.
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.