perf(export): do not allocate arrays if resource has no pending async attributes

[Samuron]

Which problem is this PR solving?

OpenTelemetry is producing too much overhead for scenarios with significant load. One of the example is BatchSpanExporter, that is allocating array of batch size only to check is it has any pending attributes.

Fixes # (issue)

Short description of the changes

Replaced map & filter with a simple loop.

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Benchmarks:

Before
create spans (10 attributes) x 689,041 ops/sec ±1.54% (92 runs sampled)
BatchSpanProcessor process span x 660,065 ops/sec ±1.08% (95 runs sampled)

After:
create spans (10 attributes) x 688,731 ops/sec ±1.15% (94 runs sampled)
BatchSpanProcessor process span x 673,550 ops/sec ±1.01% (93 runs sampled)

• Test A

Checklist:

• Followed the style guidelines of this project
• Unit tests have been added
• Documentation has been updated

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: Samuron / name: Ievgen Makukh (c5b52e2, fee30fd, a4f45c7)
• ✅ login: pichlermarc / name: Marc Pichler (670bfbb)

[Samuron]

I am not sure that the benchmarks are very representational though, they fluctuate a lot and the numbers are suspiciously close to just creating spans

[codecov]

Codecov Report

Merging #4550 (670bfbb) into main (f9a0d3e) will decrease coverage by 0.01%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4550      +/-   ##
==========================================
- Coverage   92.84%   92.84%   -0.01%     
==========================================
  Files         328      327       -1     
  Lines        9487     9471      -16     
  Branches     2035     2032       -3     
==========================================
- Hits         8808     8793      -15     
+ Misses        679      678       -1     

Files	Coverage Δ
...dk-trace-base/src/export/BatchSpanProcessorBase.ts	95.31% <100.00%> (+1.09%)	⬆️

... and 1 file with indirect coverage changes

[pichlermarc]

Thanks! Please also add a changelog entry under CHANGELOG.md (scope is sdk-trace-base, enhancement).

Ad benchmarks: I think they're not too representative in his case as they all only use a single span. If there was a benchmark with thousands of spans I think the improvement may be more visible.