[tracegen] Enable BlockOnQueueFull in OTel SDK to avoid dropped spans

[haanhvu]

Which problem is this PR solving?

[tracegen] Enable BlockOnQueueFull in OTel SDK to make spans not dropped in spikes

Short description of the changes

Call WithBlocking() from OTel SDK to enable BlockOnQueueFull=true to make spans not dropped when the queue is full.

Tested locally:

• Before the change: Spans in tracegen were dropped when generated to jaeger-collector, even with just 1m duration and 1 worker.
• With the change: Spans in tracegen were not dropped even with 1h duration and 10 workers.

[codecov]

Codecov Report

Patch coverage has no change and project coverage change: +0.01 🎉

Comparison is base (e2057fe) 97.05% compared to head (4a64ff0) 97.07%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4578      +/-   ##
==========================================
+ Coverage   97.05%   97.07%   +0.01%     
==========================================
  Files         301      301              
  Lines       17837    17837              
==========================================
+ Hits        17312    17315       +3     
+ Misses        421      419       -2     
+ Partials      104      103       -1     

Flag	Coverage Δ
unittests	97.07% <ø> (+0.01%)	⬆️

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

see 1 file with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[yurishkuro]

@haanhvu in the future, use a branch in a fork to create PRs, don't do it from main. It both breaks the CI and makes it harder for you to keep main in sync with upstream.

[haanhvu]

@haanhvu in the future, use a branch in a fork to create PRs, don't do it from main. It both breaks the CI and makes it harder for you to keep main in sync with upstream.

Yeah, I just did a super quick change on main since it's a tiny PR. Forgot that CI didn't accept it...