Http output - automatic_retries retry period

Elastic Stack Logstash
1

I'm currently using the http output plugin in order to post to an API. I noticed that the default for automatic_retries is just 1. We essentially want this to keep retrying until it succeeds, but I don't see a straightforward way of doing this. For this reason I've set the retry count to something fairly high (1000) assuming that it retries every second -- it would retry for at least 10 minutes. This seems like a bit of overkill -- but I'm looking for the following answers:

  1. How often do these retries occur? Is it just one right after the other? Or is there some "sleep" period between retries?
  2. Is there a way to have these just retry indefinitely? We're inserting messages into this API and don't want to lose any messages, so this would be the preferred method. Right now I just have any failures written off to a file and then we reprocess the fallout to be placed back to the API.
2

Logstash uses the Manticore library, which in turn in a JRuby wrapper of the Apache HttpClient library. I poked around to find something about the retry interval but didn't find anything, but that's probably where you should be looking.