opentelemetry-sdk: reduce lock contention in attributes

[codeboten]

Description

There are a few improvements here:

• set_attributes was double locking (once in set_attributes and once in __setitem__. This change updates the call to a private _setitem_locked for set_attributes once the lock has been obtained
• __iter__ doesn't use a lock if the BoundedAttributes's instantiated as immutable
• _worker_awaken was being set even when it's already set, adding a check in emit to avoid unnecessary calls

Type of change

Please delete options that are not relevant.

• Small performance improvements

How Has This Been Tested?

Ran unit tests to validate functionality was still passing and ran benchmarks including a new test to try and capture GIL contention. See the results below:

Test	Before OPS	After OPS	OPS Δ	Before (ns)	After (ns)	Latency Δ	Memory before→after
test_bounded_attribute_iterator[10]	6,900,599	7,044,875	+2.1%	144.2	144.6	+0.3%	208 → 208 (+0.0%)
test_bounded_attribute_iterator[128]	2,425,554	2,458,973	+1.4%	412.5	410.7	-0.4%	1152 → 1152 (+0.0%)
test_bounded_attribute_iterator[1]	8,008,301	8,068,857	+0.8%	124.6	125.0	+0.3%	144 → 144 (+0.0%)
test_bounded_attribute_iterator[50]	4,452,807	4,474,884	+0.5%	223.7	223.7	+0.0%	528 → 528 (+0.0%)
test_gil_contention_batch_processor[1]	392	525	+33.9%	2,319,250	1,696,584	-26.8%	n/a
test_gil_contention_batch_processor[2]	379	506	+33.4%	2,343,875	1,722,792	-26.5%	n/a
test_gil_contention_batch_processor[4]	333	435	+30.7%	2,589,709	1,818,813	-29.8%	n/a
test_gil_contention_batch_processor[8]	297	458	+54.2%	2,720,563	2,113,458	-22.3%	n/a
test_read_events[1]	158,702	165,482	+4.3%	6,250	5,792	-7.3%	964 → 964 (+0.0%)
test_read_events[10]	59,338	59,196	-0.2%	16,833	15,666	-6.9%	964 → 964 (+0.0%)
test_read_events[50]	16,001	16,134	+0.8%	62,625	60,417	-3.5%	1140 → 1140 (+0.0%)
test_read_events[128]	6,425	6,418	-0.1%	154,750	146,417	-5.4%	2372 → 2372 (+0.0%)
test_read_links[1]	176,111	177,023	+0.5%	5,541	5,375	-3.0%	964 → 964 (+0.0%)
test_read_links[10]	79,587	79,777	+0.2%	11,917	12,500	+4.9%	964 → 964 (+0.0%)
test_read_links[50]	23,526	24,213	+2.9%	41,708	41,583	-0.3%	1140 → 1140 (+0.0%)
test_read_links[128]	9,917	10,148	+2.3%	100,083	98,667	-1.4%	2372 → 2372 (+0.0%)
test_start_span_with_attributes[0]	224,978	227,978	+1.3%	4,167	4,208	+1.0%	n/a
test_start_span_with_attributes[1]	199,233	213,340	+7.1%	4,875	4,500	-7.7%	n/a
test_start_span_with_attributes[10]	140,692	144,839	+2.9%	6,792	6,625	-2.5%	n/a
test_start_span_with_attributes[50]	58,161	59,936	+3.1%	17,125	15,958	-6.8%	n/a
test_start_span_with_attributes[128]	27,675	27,575	-0.4%	36,458	35,833	-1.7%	n/a

Does This PR Require a Contrib Repo Change?

• No.

Checklist:

• Followed the style guidelines of this project
• Changelogs have been updated
• Benchmark have been added
• Documentation has been updated

[herin049]

Can't lines 284-295 be moved out of the critical section?