Empty DeleteMessageBatch error message

[jplock]

What were you trying to accomplish?

I'm using the new SQS message processor and I'm seeing a lot of these:

[ERROR] EmptyBatchRequest: An error occurred (AWS.SimpleQueueService.EmptyBatchRequest) when calling the DeleteMessageBatch operation: There should be at least one DeleteMessageBatchRequestEntry in the request.
Traceback (most recent call last):
  File "/opt/python/aws_lambda_powertools/metrics/metrics.py", line 144, in decorate
    response = lambda_handler(event, context)
  File "/opt/python/aws_lambda_powertools/tracing/tracer.py", line 282, in decorate
    response = lambda_handler(event, context)
  File "/opt/python/aws_lambda_powertools/logging/logger.py", line 246, in decorate
    return lambda_handler(event, context)
  File "/opt/python/aws_lambda_powertools/middleware_factory/factory.py", line 130, in wrapper
    response = middleware()
  File "/opt/python/aws_lambda_powertools/utilities/batch/sqs.py", line 179, in sqs_batch_processor
    processor.process()
  File "/opt/python/aws_lambda_powertools/utilities/batch/base.py", line 58, in __exit__
    self._clean()
  File "/opt/python/aws_lambda_powertools/utilities/batch/sqs.py", line 118, in _clean
    delete_message_response = self.client.delete_message_batch(QueueUrl=queue_url, Entries=entries_to_remove)
  File "/opt/python/botocore/client.py", line 337, in _api_call
    return self._make_api_call(operation_name, kwargs)
  File "/opt/python/aws_xray_sdk/ext/botocore/patch.py", line 38, in _xray_traced_botocore
    return xray_recorder.record_subsegment(
  File "/opt/python/aws_xray_sdk/core/recorder.py", line 435, in record_subsegment
    return_value = wrapped(*args, **kwargs)
  File "/opt/python/botocore/client.py", line 656, in _make_api_call
    raise error_class(parsed_response, operation_name)

in my logs.

Expected Behavior

No errors in logs about empty messages.

Current Behavior

Errors in logs about empty messages when calling SQS batch delete message.

Possible Solution

Maybe we could add a simple check around entries_to_remove to ensure it contains any entries before calling
delete_message_batch?

Steps to Reproduce (for bugs)

1. have only a single message in the queue
2. set the BatchSize to one so the Lambda only receives a single record
3. Have the record_handler throw an exception

Environment

• Powertools version used: 1.5.0
• Packaging format (Layers, PyPi): Layers
• AWS Lambda function runtime: Python 3.8
• Debugging logs

How to enable debug mode**

# paste logs here

[heitorlessa]

hey @jplock - Thanks for reporting it!

I was nearly sure this was taken care of here:

• Return if no failed messages: https://github.com/awslabs/aws-lambda-powertools-python/blob/b654ca68a5418af6f6728670d2851034adf79bcb/aws_lambda_powertools/utilities/batch/sqs.py#L111
• Test for no failed messages to be deleted: https://github.com/awslabs/aws-lambda-powertools-python/blob/b654ca68a5418af6f6728670d2851034adf79bcb/tests/functional/test_utilities_batch.py#L92

@cakepietoast could you have a look when you get a chance, please?

[to-mc]

Looks like we'd missed testing the case where there are only failures. PR #170 adds a test for this and fixes the issue.

[to-mc]

This is merged, will be fixed when we release 1.6.0

[jplock]

Thanks for the quick turnaround!

[to-mc]

@jplock this should be fixed now - please confirm when you have a chance to test and we'll close the issue!

[jplock]

Looks good, thanks everyone!