BatchProcessor: pattern/best practices for passing additional parameters from lambda_handler() to record_handler()?

[cmurtaugh]

Hi --

I'm following the documentation for batch-processing SQS messages, and it works well. I'd like to know if there are any examples / best practices for how to pass additional information from the lambda_handler to the record_handler. For example, I want to retrieve some SSM parameters in the lambda_handler code and make them available to the record_handler code. Is there a good way to do this?

Thanks!
--Colin

[goutamadwant]

BatchProcessor only injects the record (and, when requested by name, lambda_context); it does not have a general dependency-injection mechanism. A closure or functools.partial is therefore a clean way to bind invocation-scoped values:

from functools import partial
from aws_lambda_powertools.utilities import parameters
from aws_lambda_powertools.utilities.batch import process_partial_response

def record_handler(record, *, settings):
    # use settings while processing this record
    ...

def lambda_handler(event, context):
    settings = parameters.get_parameters("/my/service/", max_age=60)

    return process_partial_response(
        event=event,
        record_handler=partial(record_handler, settings=settings),
        processor=processor,
        context=context,
    )

This fetches once for the invocation, not once per record, and keeps invocation-specific state out of globals. If the values may safely be reused across warm invocations, Powertools Parameters already has an in-memory TTL cache, so a module-level provider plus an appropriate max_age is usually preferable to hand-rolled caching.

The important part is that the bound callable still accepts record; the processor invokes it as handler(record=data).