Problem with chunks/times?

[digitalsanity]

There is an issue with fluent-bit storage.max_chunks_up > 1 (default is 128) and larger streams -- the chunks all come in with the same timestamp (microseconds difference) so you end up overwriting the same object in S3 and losing data.

{"fields":{},"level":"info","timestamp":"2020-03-04T07:17:43.618121481Z","message":"Successfully wrote s3://ingest.test/new/test6/20200304/7/20200304071743_ingest.test.log"}
{"fields":{},"level":"info","timestamp":"2020-03-04T07:17:43.689532265Z","message":"Successfully wrote s3://ingest.test/new/test6/20200304/7/20200304071743_ingest.test.log"}
{"fields":{},"level":"info","timestamp":"2020-03-04T07:17:43.785056959Z","message":"Successfully wrote s3://ingest.test/new/test6/20200304/7/20200304071743_ingest.test.log"}
{"fields":{},"level":"info","timestamp":"2020-03-04T07:17:43.86481822Z","message":"Successfully wrote s3://ingest.test/new/test6/20200304/7/20200304071743_ingest.test.log"}

Fix is changing timestamp in GenerateObjectKey to:
timestamp := t.Format("20060102150405.0000")