fix(net.peer.port): net.peer.port needs to be a number not a string

[esara]

I noticed that a number of contrib instrumentation populated the net.peer.port as a string parsed out of the connection string. This not only violates the semantic convensations, but in same cases causes and exception the the span attributes are parsed. https://github.com/open-telemetry/opentelemetry-specification/blob/v1.20.0/specification/trace/semantic_conventions/database.md#connection-level-attributes

Which problem is this PR solving?

This is related to #781
More specifically, there is an inconsistency, where the port variable is a string in multiple instrumentation (memcached, mongodb, mysql/mysql2, mongodb) for example
https://github.com/open-telemetry/opentelemetry-js-contrib/blob/main/plugins/node/opentelemetry-instrumentation-mongodb/src/instrumentation.ts#L802
compared to other instrumentation, where is correctly defined as an int (cassandra and others)
https://github.com/open-telemetry/opentelemetry-js-contrib/blob/main/plugins/node/opentelemetry-instrumentation-cassandra/src/instrumentation.ts#L179

Short description of the changes

Making sure that the net.peer.port attribute is a number instead of a string

BEGIN_COMMIT_OVERRIDE
fix(instrumentation-memcached): parse attribute value for net.peer.port to a number
fix(instrumentation-mongodb): parse attribute value for net.peer.port to a number
fix(instrumentation-mysql): parse attribute value for net.peer.port to a number
fix(instrumentation-mysql2): parse attribute value for net.peer.port to a number
END_COMMIT_OVERRIDE

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: esara / name: Endre Sara (247a144, 2768981)

[pichlermarc]

Looks good! 🙂
Thanks for taking care of this. 🙂

[codecov]

Codecov Report

Merging #1916 (2768981) into main (5b0b67f) will decrease coverage by 0.03%.
The diff coverage is 93.33%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1916      +/-   ##
==========================================
- Coverage   91.02%   91.00%   -0.03%     
==========================================
  Files         146      146              
  Lines        7478     7491      +13     
  Branches     1497     1502       +5     
==========================================
+ Hits         6807     6817      +10     
- Misses        671      674       +3     

Files	Coverage Δ
...try-instrumentation-mongodb/src/instrumentation.ts	53.27% <100.00%> (+0.42%)	⬆️
...e/opentelemetry-instrumentation-mysql/src/utils.ts	97.72% <100.00%> (-2.28%)	⬇️
.../opentelemetry-instrumentation-mysql2/src/utils.ts	96.96% <100.00%> (-3.04%)	⬇️
...entelemetry-instrumentation-memcached/src/utils.ts	80.00% <80.00%> (-1.25%)	⬇️

[pichlermarc]

Ah looks like the tests also need to be updated (they assert that the values are strings, instead of number)

[esara]

Ah looks like the tests also need to be updated (they assert that the values are strings, instead of number)

@pichlermarc let me know if the test look better now

[pichlermarc]

@esara, please don't force-push. It makes it hard to review the PR because the history disappears every time I check back 😅

(also makes it hard to see if the tests were passing as I have to dig into old workflow runs to find the run that was associated with the commit before the force-push)