Fix histogram collect benchmark

[fraillt]

Changes

There are two problems with current benchmark for histogram collection:

• it simply doesn't work...

cargo bench --bench metric Histogram/Collect
   Compiling opentelemetry_sdk v0.26.0 (/home/fraillt/Projects/opentelemetry-rust/opentelemetry-sdk)
    Finished `bench` profile [optimized + debuginfo] target(s) in 5.46s
     Running benches/metric.rs (target/release/deps/metric-36200279505f7f44)
Gnuplot not found, using plotters backend
Benchmarking Histogram/CollectOne: Warming up for 3.0000 sthread 'main' panicked at opentelemetry-sdk/benches/metric.rs:398:36:
index out of bounds: the len is 0 but the index is 0
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

• it only test case where measurements has 0 attributes

This revision fixes first problem, and broadens scope of measuring more combinations:

    for metrics_count in [1, 20] {
        for (attribute_sets_count, attribs_count) in [(1, 0), (1, 10), (500, 10)] {
        ...
        }
    }

Please provide a brief description of the changes here.

Merge requirement checklist

• CONTRIBUTING guidelines followed
• Unit tests added/updated (if applicable)
• Appropriate CHANGELOG.md files updated for non-trivial, user-facing changes
• Changes in public API reviewed (if applicable)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 79.4%. Comparing base (fcd7cae) to head (1e5fff3).

Additional details and impacted files

@@           Coverage Diff           @@
##            main   #2268     +/-   ##
=======================================
- Coverage   79.4%   79.4%   -0.1%     
=======================================
  Files        121     121             
  Lines      20981   20981             
=======================================
- Hits       16673   16671      -2     
- Misses      4308    4310      +2     

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