feat: Allowed undici error reporting to be disabled with feature flag undici_error_tracking

[Voziv]

Description

When fetch in node 20+ is wrapped in a try/catch the error that is thrown even if the user doesn't rethrow the error. At best this means errors being reported that are not errors, at worst it means errors are being reported twice.

This adds the feature flag undici_error_tracking as suggested by @bizob2828 to allow disabling error capture via undici message channels.

How to Test

See test in the files changed. I've caught an undici error and ensured that when the flag is turned off the transaction error array is empty.

Related Issues

Fixes #2954

[CLAassistant]

All committers have signed the CLA.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[bizob2828]

Looks good. I'm going to commit a documentation update and test tweak but thanks for adding @Voziv !

[bizob2828]

i cannot push to your branch, so i suggested a change and I'll follow up with doc update

[bizob2828]

thanks for handling the back and forth @Voziv, LGTM. Just a heads up, we will probably release this next week.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.30%. Comparing base (bc714cf) to head (6db3786).
Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2956      +/-   ##
==========================================
- Coverage   97.35%   97.30%   -0.05%     
==========================================
  Files         317      317              
  Lines       48690    48692       +2     
==========================================
- Hits        47402    47382      -20     
- Misses       1288     1310      +22     

Flag	Coverage Δ
integration-tests-cjs-18.x	72.34% <100.00%> (+<0.01%)	⬆️
integration-tests-cjs-20.x	72.34% <100.00%> (+<0.01%)	⬆️
integration-tests-cjs-22.x	72.55% <100.00%> (-0.01%)	⬇️
integration-tests-esm-18.x	49.96% <50.00%> (ø)
integration-tests-esm-20.x	49.97% <50.00%> (+<0.01%)	⬆️
integration-tests-esm-22.x	50.02% <50.00%> (ø)
unit-tests-18.x	89.03% <100.00%> (+<0.01%)	⬆️
unit-tests-20.x	89.03% <100.00%> (+<0.01%)	⬆️
unit-tests-22.x	89.03% <100.00%> (+<0.01%)	⬆️
versioned-tests-18.x	79.58% <100.00%> (-0.19%)	⬇️
versioned-tests-20.x	79.58% <100.00%> (-0.19%)	⬇️
versioned-tests-22.x	79.59% <100.00%> (-0.19%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

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