fix(instr-mysql2): Fix mysql2 instrumentation connection prototype

[david-luna]

Which problem is this PR solving?

mysql2@3.11.5 introduced a internal refactoring to solve circular dependencies within the package.
ref: sidorares/node-mysql2#3081

The changes include a new BaseConnection class which contains the methods we need to patch (query, execute). The Connection class is extending it therefore the instrumentation gets the wrong prototype when patching.

Fixes: #2572

Short description of the changes

• bump mysql2 to v3.11.5
• added a function to extract the proper prototype when patching/unpatching

[codecov]

Codecov Report

Attention: Patch coverage is 85.71429% with 1 line in your changes missing coverage. Please review.

Project coverage is 90.75%. Comparing base (8bfa21f) to head (b5b72ec).

Files with missing lines	Patch %	Lines
...etry-instrumentation-mysql2/src/instrumentation.ts	85.71%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2578      +/-   ##
==========================================
- Coverage   90.75%   90.75%   -0.01%     
==========================================
  Files         169      169              
  Lines        8018     8023       +5     
  Branches     1632     1633       +1     
==========================================
+ Hits         7277     7281       +4     
- Misses        741      742       +1     

Files with missing lines	Coverage Δ
...etry-instrumentation-mysql2/src/instrumentation.ts	93.75% <85.71%> (-0.92%)	⬇️

[pichlermarc]

ah I also just finished fixing this on my working copy before heading out to lunch, looks like you beat my by half an hour 🙂

Looks good, thanks 🙌 Let's wait for TAV to complete and then we should be good to go 🙂

[pichlermarc]

Huh, weird - tests in TAV seem to be still failing but it's not yet possible to see which versions they're failing at. 🤔

I think I missed something that keeps failing tests

[david-luna]

ah I also just finished fixing this on my working copy before heading out to lunch, looks like you beat my by half an hour 🙂

Looks good, thanks 🙌 Let's wait for TAV to complete and then we should be good to go 🙂

Forgot to link the issue in the PR, sorry. TAV tests seem to be stuck so maybe my fix is not solving all issues :(

[pichlermarc]

ah I also just finished fixing this on my working copy before heading out to lunch, looks like you beat my by half an hour 🙂
Looks good, thanks 🙌 Let's wait for TAV to complete and then we should be good to go 🙂

Forgot to link the issue in the PR, sorry. TAV tests seem to be stuck so maybe my fix is not solving all issues :(

No worries 🙂

I've also pushed my change (it's very similar to yours) - the only difference I can tell is that I'm checking that the base is BaseConnection. Let's see if that passes the TAV script 🤔

#2579

[pichlermarc]

Ah, I think in older versions you get EventEmitter as a base. That may be why the TAV tests are failing.

[david-luna]

I've reverted the check to make it more explicit. It will return the base class prototype if it has the query function (execute goes together). Otherwise it falls back to the Connection prototype.

Note: I had trouble running it locally since docker image mysql:5.7 has no support for linux/arm64/v8 platform. Is it possible to update the image version in @opentelemetry/contrib-test-utils?

[david-luna]

closing this one in favor of #2579