opentelemetry-sdk: drop NaN measurements at instrument level to prevent aggregation poisoning

[RudraDudhat2509]

Summary

• NaN measurements passed to metric instruments (e.g. counter.add(float("nan"))) were silently folded into running aggregations, permanently poisoning the sum, all subsequent exports emitted nan with no recovery path
• Root cause: the existing if amount < 0 guard on Counter and Histogram is bypassed by NaN because all IEEE 754 comparisons against NaN return False
• Fix: add a math.isnan() check at the instrument level (same layer as the existing negative-value guard) for all sync instruments (Counter, UpDownCounter, Histogram, Gauge) and in the async callback path, matching the approach taken in the Java SDK (open-telemetry/opentelemetry-java#5846, fixed in PR #5859)

Closes #5330

What was tested

• Added test_add_nan to TestCounter and TestUpDownCounter
• Added test_record_nan to TestHistogram
• Added test_set_nan to TestGauge
• Added test_nan_callback_value_is_dropped to TestObservableGauge (verifies NaN from async callbacks is skipped, valid measurements still pass through)
• All 30 instrument tests pass (python -m pytest opentelemetry-sdk/tests/metrics/test_instrument.py)
• Verified end-to-end: counter.add(1) → add(NaN) → add(5) now yields sum = 6, not nan

[RudraDudhat2509]

here is a ss of the fix for the teams reference

[herin049]

Changelog fragment number should match the PR number

[RudraDudhat2509]

Recent commit addresses the same, thx for pointing it out !