Batches size now affects sending size accurately. Use extensions to …

[adamhathcock]

…help test this

This was fixed but had a regression

This caused a single request to be too big sometimes and it was noticeable on large models. This fix says anything larger than 25 is 25 so an individual request won't be (much) larger than 25 megs. There are tests now to prevent this being reverted again.

[linear]

CXPLA-166 V3 Large models 413

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 70.93%. Comparing base (e5a0915) to head (1d80b44).
Report is 1 commits behind head on dev.

Additional details and impacted files

@@             Coverage Diff             @@
##              dev     #230       +/-   ##
===========================================
+ Coverage   59.67%   70.93%   +11.25%     
===========================================
  Files         262      279       +17     
  Lines       10489    11256      +767     
  Branches     1107     1109        +2     
===========================================
+ Hits         6259     7984     +1725     
+ Misses       3969     2919     -1050     
- Partials      261      353       +92     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[JR-Morgan]

• Rhino Sends - Confirmed fixes large send issue

• I would like to debug some more to measure what size of batches we are now sending
  Confirmed, not a single batch was over 20MiB

[adamhathcock]

25 meg or less if nothing is larger than 25 meg.

Batches could be over that if a single message is over 25 megs 😬

[adamhathcock]

this adds logging as a warning for large requests: #232