Lambda Operators & Hooks

The Operator

from airflow.providers.amazon.aws.operators.lambda_function import LambdaInvokeFunctionOperator

invoke_fn = LambdaInvokeFunctionOperator(
    task_id="invoke_processing_function",
    function_name="sales-data-validator",
    payload='{"table": "orders", "date": "2026-09-04"}',
    aws_conn_id="aws_default",
)

LambdaInvokeFunctionOperator supports both invocation types:

Run for real, invoking a live Lambda function and passing its response downstream via XCom:

Figure — a genuine synchronous Lambda invocation; the second task receives the actual response payload through .output.

The Hook Directly

from airflow.providers.amazon.aws.hooks.lambda_function import LambdaHook

def invoke_with_retry_logic(**context):
    hook = LambdaHook(aws_conn_id="aws_default")
    response = hook.invoke_lambda(
        function_name="sales-data-validator",
        payload='{"table": "orders"}',
    )
    return response["Payload"].read().decode()

Reach for the Hook directly when you need custom error handling around the invocation (e.g., a specific retry policy for TooManyRequestsException throttling) that the operator's built-in retries don't cover.

This content is reserved for Premium Members.