Support JSON Lines in InfluxDB Edge

[mgattozzi]

As part of #24616 we added JSON support, however, this has the problem of requiring us to buffer all of the data into memory. Given a sufficiently large query this would cause a possible OOM kill and so we should offer JSON Lines where each line is a row from the returned query. This would allow us to stream the data from a SendableRecordBatchStream as we get it, to the user, and avoid buffering too much data into memory.

Steps:

1. Create a JsonLinesFormatter in the influxdb3_server/src/http.rs file that uses the JsonFormat trait to set the output to follow JSON lines e.g. instead of the JSON:

[ { "foo": "bar" }, {"foo": "baz" } ]

it should be output as:

{ "foo": "bar" }
{ "foo": "baz" }

where each record is on it's own line without the []

This should just be writing a '\n' char into the writer in the end_row function and the rest should just need to be stubbed out with Ok(())

2. Using the Writer and the JsonLinesFormatter write the data into a streaming body in a separate task, while returning the Response for the API in the query routes