Mirror influxdb api for large payloads in influx_listener input

[glinton]

Addresses comments in #2280

An attempt was made to add a content-length header to the influxdb output, but it was decided that would cost more than it was worth (read the body to count bytes, then allocate a new reader to use in the request)

[danielnelson]

The old code is too tricky, I don't feel comfortable that it works. We also do a bunch of work so that we can apply a line length limit, a feature that InfluxDB doesn't even appear to have.

Let's remove the ability to limit lines, accepting the option but warning when it is set, and only retain max body limit. This will allow us to pass the full body into the parser and we can use much simpler processing.

This will be needed anyway to solve #6124.

[glinton]

Closing in favor of http_listener rewrite referred to here