Missing event writing to S3 when sending 200 concurrent requests

[lrivera-verimatrix]

We are experiencing data loss when writing our events to S3 using the below configuration. We do see everything in STDOUT, and to the FILE, but not in S3. We do not really know what is wrong, and it would be very helpful if someone could point out what might be wrong here. Or is this a bug?

This is using and EC2 with the Amazon version of td-agent: td-agent-3.6.0-0.amazon2.x86_64. The S3 is in the same region (us-east-1), so I do not believe this is a connection or bandwidth issue. We can send 100 concurrent requests and everything lads in S3, but if we go to 200, the S3 one does not have everything.

<match nTitleMe*.*>
@type copy

@type file
path /ntm/logs
compress gzip

timekey 1m
timekey_use_utc true
timekey_wait 1m



@type s3

s3_bucket iad-sb-ntitleme-log-aggr
s3_region us-east-1
path ${region}/%Y/%m/%d/%H/${logtype}.
time_slice_format %Y%m%d:%H%M%S
s3_object_key_format %{path}%{time_slice}_%{index}.ip-10-232-5-127.ec2.internal.%{file_extension}

<buffer logtype,time>
  @type file
  path /mnt/volume/td-agent/buffer/s3/nTitleMeJsonEvents
  timekey 5m
  timekey_wait 1m
  timekey_use_utc true
  chunk_limit_size 256m
  flush_thread_count 8
</buffer>

@type stdout

[github-actions]

This issue has been automatically marked as stale because it has been open 90 days with no activity. Remove stale label or comment or this issue will be closed in 30 days

[github-actions]

This issue was automatically closed because of stale in 30 days