Denis

> @rdner have you seen this https://github.com/elastic/beats/actions/runs/8813847492/job/24193021847?pr=38213 again? @gizas Looks like something is wrong with the Ubuntu itself, it failed on buildkite too. I think we should try to restart,...

@elastic/obs-infraobs-integrations can someone please have a look?

I don't recall us changing anything related to index creation lately. Also, the fact it started affecting multiple versions at once suggests it's either a backported change or something else...

The condition here https://github.com/elastic/beats/blob/17416f9821e515f0b59e54650a860c40e9d99689/filebeat/input/filestream/fswatch.go#L160 Requires the file size to remain the same. If the file size has changed the file won't be resent, instead it's treated as modified and only...

Need some input from the product team, pending.

Introducing a new value for the setting with a different type ("string" vs "boolean") would make the new config incompatible with the previous versions of Filebeat. Are we fine with...

> This appears to be called during the keep alive session maintenance every time data is read from the buffer? The last data in the buffer is the 200 OK...

I think we should find the places where those "keep alive requests" happen and send them without the logger connection wrapper at all. We should not care about their success...

> This appears to be called during the keep alive session maintenance every time data is read from the buffer? The last data in the buffer is the 200 OK...

Just to summarize my thoughts on this issue: 1. We should definitely keep the log wrapper, we should not consider removing it 2. The log message must stay because it's...