Use Arc to store reference to LoggerProvider from Logger

[bIgBV]

Fixes #
Design discussion issue (if applicable) #1209

Changes

Switch from using a Weak reference to LoggerProvider from within the Logger to using an Arc.

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)

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: bIgBV / name: Bhargav (2bc0ef1, 2cb8ed2, 2ff86bd)

[bIgBV]

I wanted to get draft PR out, but I still want to validate that the change doesn't cause any subtle breakage.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (f31c4dd) 60.8% compared to head (2ff86bd) 60.8%.

Additional details and impacted files

@@          Coverage Diff          @@
##            main   #1446   +/-   ##
=====================================
  Coverage   60.8%   60.8%           
=====================================
  Files        146     146           
  Lines      19230   19230           
=====================================
  Hits       11702   11702           
  Misses      7528    7528           

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

[cijothomas]

As discussed in the community meeting today, we'll first add test coverage, targeted towards shutdown scenarios to have more confidence in this. Also will try this for span as well, where the benefits are even higher.

[bIgBV]

Ran the stress tests on my WSL system and didn't get a huge boost:

On this branch:

Number of threads: 12
Throughput: 10,784,000 iterations/sec
Throughput: 10,968,400 iterations/sec
Throughput: 10,636,200 iterations/sec
Throughput: 10,826,800 iterations/sec
Throughput: 10,634,800 iterations/sec
Throughput: 10,761,200 iterations/sec
Throughput: 10,676,800 iterations/sec
Throughput: 10,725,600 iterations/sec
Throughput: 11,487,600 iterations/sec
Throughput: 11,224,600 iterations/sec
Throughput: 10,704,400 iterations/sec
Throughput: 10,658,600 iterations/sec
Throughput: 10,730,200 iterations/sec
Throughput: 11,040,200 iterations/sec
Throughput: 10,574,400 iterations/sec
Throughput: 10,751,400 iterations/sec
Throughput: 10,721,600 iterations/sec
Throughput: 10,847,800 iterations/sec
Throughput: 10,532,800 iterations/sec
Throughput: 10,906,000 iterations/sec
Throughput: 10,960,400 iterations/sec
Throughput: 11,012,400 iterations/sec
Throughput: 10,615,400 iterations/sec
Throughput: 10,689,400 iterations/sec
Throughput: 10,802,600 iterations/sec
Throughput: 10,684,000 iterations/sec
Throughput: 10,981,200 iterations/sec
Throughput: 10,748,800 iterations/sec
Throughput: 11,100,400 iterations/sec
Throughput: 10,765,000 iterations/sec
Throughput: 10,762,200 iterations/sec
Throughput: 10,770,200 iterations/sec
Throughput: 10,589,800 iterations/sec
Throughput: 10,953,200 iterations/sec
Throughput: 10,502,000 iterations/sec
Throughput: 10,806,400 iterations/sec
Throughput: 10,971,200 iterations/sec
Throughput: 10,783,200 iterations/sec
Throughput: 10,805,800 iterations/sec
Throughput: 10,761,200 iterations/sec
Throughput: 10,649,800 iterations/sec
Throughput: 10,718,600 iterations/sec
Throughput: 10,613,600 iterations/sec
Throughput: 10,501,600 iterations/sec
Throughput: 10,660,800 iterations/sec
Throughput: 10,661,000 iterations/sec
Throughput: 10,980,600 iterations/sec
Throughput: 10,955,800 iterations/sec
Throughput: 10,936,000 iterations/sec
Throughput: 10,927,600 iterations/sec

Avg: 10,797,272
Median: 10,761,700

On main:

Number of threads: 12
Throughput: 9,412,200 iterations/sec
Throughput: 8,960,600 iterations/sec
Throughput: 8,911,000 iterations/sec
Throughput: 8,826,800 iterations/sec
Throughput: 9,044,800 iterations/sec
Throughput: 9,105,000 iterations/sec
Throughput: 9,358,800 iterations/sec
Throughput: 9,207,400 iterations/sec
Throughput: 9,128,600 iterations/sec
Throughput: 9,042,000 iterations/sec
Throughput: 9,356,600 iterations/sec
Throughput: 9,074,600 iterations/sec
Throughput: 9,392,200 iterations/sec
Throughput: 9,186,400 iterations/sec
Throughput: 9,903,600 iterations/sec
Throughput: 9,424,800 iterations/sec
Throughput: 9,000,000 iterations/sec
Throughput: 9,317,200 iterations/sec
Throughput: 9,029,600 iterations/sec
Throughput: 9,364,800 iterations/sec
Throughput: 8,980,000 iterations/sec
Throughput: 9,043,600 iterations/sec
Throughput: 8,900,400 iterations/sec
Throughput: 9,235,800 iterations/sec
Throughput: 8,994,400 iterations/sec
Throughput: 9,101,800 iterations/sec
Throughput: 8,968,400 iterations/sec
Throughput: 9,124,600 iterations/sec
Throughput: 9,067,800 iterations/sec
Throughput: 8,962,800 iterations/sec
Throughput: 8,922,800 iterations/sec
Throughput: 8,915,000 iterations/sec
Throughput: 9,121,200 iterations/sec
Throughput: 8,973,400 iterations/sec
Throughput: 9,315,800 iterations/sec
Throughput: 9,107,600 iterations/sec
Throughput: 9,565,800 iterations/sec
Throughput: 9,081,800 iterations/sec
Throughput: 8,814,200 iterations/sec
Throughput: 8,894,800 iterations/sec
Throughput: 9,355,400 iterations/sec
Throughput: 9,143,000 iterations/sec
Throughput: 9,100,200 iterations/sec
Throughput: 9,227,800 iterations/sec
Throughput: 9,019,600 iterations/sec
Throughput: 9,529,800 iterations/sec
Throughput: 9,417,600 iterations/sec
Throughput: 8,983,600 iterations/sec
Throughput: 9,064,200 iterations/sec
Throughput: 9,095,400 iterations/sec
Throughput: 9,360,200 iterations/sec
Throughput: 8,971,400 iterations/sec
Throughput: 9,077,000 iterations/sec
Throughput: 9,086,400 iterations/sec
Throughput: 9,015,200 iterations/sec
Throughput: 9,390,600 iterations/sec
Throughput: 7,884,000 iterations/sec

Avg: 9,120,357.894
Median: 9,086,400

The interesting thing is, I ran smaller tests on my m2 macbook pro and the difference was much larger. I'll put those numbers up later today.

[bIgBV]

From my m2 MBP

on the current branch:

Number of threads: 10
Throughput: 6,000,000 iterations/sec
Throughput: 6,033,000 iterations/sec
Throughput: 6,015,800 iterations/sec
Throughput: 6,009,400 iterations/sec
Throughput: 6,028,200 iterations/sec
Throughput: 6,013,400 iterations/sec
Throughput: 6,011,400 iterations/sec
Throughput: 5,993,200 iterations/sec
Throughput: 6,003,600 iterations/sec
Throughput: 5,994,000 iterations/sec
Throughput: 6,005,600 iterations/sec
Throughput: 5,977,000 iterations/sec
Throughput: 6,007,800 iterations/sec
Throughput: 6,015,600 iterations/sec

vs main

Number of threads: 10
Throughput: 3,491,000 iterations/sec
Throughput: 3,481,000 iterations/sec
Throughput: 3,494,200 iterations/sec
Throughput: 3,485,600 iterations/sec
Throughput: 3,511,600 iterations/sec
Throughput: 3,530,200 iterations/sec
Throughput: 3,520,600 iterations/sec
Throughput: 3,592,800 iterations/sec
Throughput: 3,540,400 iterations/sec
Throughput: 3,522,000 iterations/sec
Throughput: 3,724,400 iterations/sec
Throughput: 3,668,800 iterations/sec
Throughput: 3,526,800 iterations/sec
Throughput: 3,520,200 iterations/sec
Throughput: 3,648,200 iterations/sec
Throughput: 3,576,200 iterations/sec
Throughput: 3,646,600 iterations/sec
Throughput: 3,585,400 iterations/sec
Throughput: 3,541,800 iterations/sec
Throughput: 3,514,400 iterations/sec
Throughput: 3,520,600 iterations/sec
Throughput: 3,582,800 iterations/sec
Throughput: 3,574,200 iterations/sec

[bIgBV]

Looks like the git history is messed up, will fix it tomorrow.

[lalitb]

Ran the stress tests on my WSL system and didn't get a huge boost:

@bIgBV It could be the compiler optimizations for Mac and WSL that we are seeing the difference. Can you confirm if you tested this in release mode:

<path-to-otel-rust>/stress$ cargo run --bin logs --release

[bIgBV]

@lalitb I can confirm that the stress test was done in release mode on both systems.

[bIgBV]

Closing this PR in favor of #1455