fix(cluster): reduce lock contention on cluster initialization

[crenshaw-dev]

c.newResource calls c.populateResourceInfoHandler which, in Argo CD, may do nontrivial work: https://github.com/argoproj/argo-cd/blob/75def4f2df3e27892292b8020bfb9100a2784105/controller/cache/cache.go#L532-L560

As far as I can tell, c.newResource doesn't do any work which requires the lock.

This is the benchmark before/after:

Benchmark_sync-16    	     100	  10186413 ns/op
Benchmark_sync-16    	     212	   5531238 ns/op

Under the benchmark's conditions, cluster initialization is about twice as fast.

Tests pass in Argo CD: argoproj/argo-cd#21666

[sonarqubecloud]

Quality Gate passed

Issues
1 New issue
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 54.44%. Comparing base (8849c3f) to head (90fece3).
Report is 6 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #660      +/-   ##
==========================================
+ Coverage   54.26%   54.44%   +0.17%     
==========================================
  Files          64       64              
  Lines        6164     6271     +107     
==========================================
+ Hits         3345     3414      +69     
- Misses       2549     2583      +34     
- Partials      270      274       +4     

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

[agaudreault]

LGTM. I dont see how this would cause a deadlock because "newResources" does not assume it has been locked and the populateResourceInfoHandler shouldnt already not perform operation on the cache that would require a lock.