Attempt to write metric tests with inmemoryexp

[cijothomas]

Attempting to write Metrics test to ensure SDK is doing the aggregation as expected. This is full of "hacks" to unblock.! Need to spend some time to understand the inner workings to fix this - If anyone can help, really appreciate it! Thanks @TommyCpp who suggested the fix for flush being stuck!

Such tests are very critical to make sure the aggregation is correct, especially considering we are considering big changes like the ones in this PR - #1379

[codecov]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (9e2e3db) 55.6% compared to head (364d736) 57.1%.
Report is 1 commits behind head on main.

Files	Patch %	Lines
opentelemetry-sdk/src/metrics/mod.rs	98.7%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##            main   #1382     +/-   ##
=======================================
+ Coverage   55.6%   57.1%   +1.4%     
=======================================
  Files        145     146      +1     
  Lines      18022   18099     +77     
=======================================
+ Hits       10037   10350    +313     
+ Misses      7985    7749    -236     

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

[hdost]

Looks like a winner to me 🤘 one step closer to stability.

[TommyCpp]

Good first step 👍 we can try table-driven unit tests when adding more test cases

[lalitb]

LGTM. Nicely done :)

[lalitb]

nit - attributes can be defined once, and reused - something like (not tested):

let key_value1 = KeyValue::new("key1", "value1"); 
let key_value2 = KeyValue::new("key1", "value2");

for _ in 0..5 {
    counter.add(1, &[key_value1]);  // or &[key_value1.clone()] if metrics SDK takes ownership

for _ in 0..3 {
    counter.add(1, &[key_value2]);
}

[cijothomas]

This is far from ideal, but merging as is, give this works, and can be used to quickly validate that aggregations/merges are done correctly, especially when working on PRs like #1379

Expecting to improve such tests in follow ups. Thanks everyone for helping review!