Benchmark test in Typescript

[martinkuba]

This is the same as #4105, but written in Typescript.

[codecov]

Codecov Report

Merging #4143 (59b5529) into main (cbc5c52) will decrease coverage by 0.48%.
The diff coverage is n/a.

❗ Current head 59b5529 differs from pull request most recent head 9ef6acc. Consider uploading reports for the commit 9ef6acc to get more accurate results

@@            Coverage Diff             @@
##             main    #4143      +/-   ##
==========================================
- Coverage   92.72%   92.24%   -0.48%     
==========================================
  Files         298      326      +28     
  Lines        8297     9299    +1002     
  Branches     1728     1971     +243     
==========================================
+ Hits         7693     8578     +885     
- Misses        604      721     +117     

[pichlermarc]

Thanks for taking the time to create this PR. I think prefer this one. 🙂

cc @open-telemetry/javascript-approvers thoughts? I think this is here a bit more seamless with our other code also written typescript 🙂

[dyladan]

I think I might prefer JS actually in this context. If we update our version of typescript it might change the benchmarks. It's ok if it changes the performance of the library because that's what we're measuring, but we want the thing we use to measure performance to be as stable as possible. If we update TS and the benchmark gets 5% slower is that because the library got slower or because the benchmark got slower (admittedly a synthetic example, but i think it gets the point across)?

[pichlermarc]

I think I might prefer JS actually in this context. If we update our version of typescript it might change the benchmarks. It's ok if it changes the performance of the library because that's what we're measuring, but we want the thing we use to measure performance to be as stable as possible. If we update TS and the benchmark gets 5% slower is that because the library got slower or because the benchmark got slower (admittedly a synthetic example, but i think it gets the point across)?

Yep, that makes sense. @martinkuba thanks for putting in the effort to open this PR and sorry for the churn - let's go with #4105