Error sending empty batch on shutdown

[zl4bv]

I see the following error for each configured usage of the kinesis firehose plugin whenever fluent-bit shuts down.

It looks like the plugin is trying to flush any remaining records and attempts to send an empty batch. Not sure if my configuration needs to be adjusted or perhaps some additional validation in the plugin is needed to ensure a batch is only sent if it contains at least one record.

[engine] caught signal (SIGTERM)
[2020/06/30 06:11:45] [ info] [input] pausing tail.0
[2020/06/30 06:11:45] [ info] [input] pausing tail.1
[2020/06/30 06:11:45] [ info] [input] pausing tail.2
[2020/06/30 06:11:45] [ info] [input] pausing tail.3
[2020/06/30 06:11:45] [ info] [input] pausing tail.4
[2020/06/30 06:11:45] [ info] [input] pausing tail.5
[2020/06/30 06:11:45] [ warn] [engine] service will stop in 5 seconds
[engine] caught signal (SIGTERM)
[2020/06/30 06:11:45] [ warn] [engine] service will stop in 5 seconds
[2020/06/30 06:11:50] [ info] [engine] service stopped
time="2020-06-30T06:11:50Z" level=error msg="[firehose 0] PutRecordBatch failed with InvalidParameter: 1 validation error(s) found.\n- minimum field size of 1, PutRecordBatchInput.Records.\n"
time="2020-06-30T06:11:50Z" level=error msg="[firehose 0] InvalidParameter: 1 validation error(s) found.\n- minimum field size of 1, PutRecordBatchInput.Records.\n\n"

I'm running aws-for-fluent-bit 2.3.1 which is bundled with Fluent Bit 1.4.5 and Amazon Kinesis Firehose for Fluent Bit 1.2.1.