fix: Fixed event matcher to focus on properties specific to web requests (v1/ALB and v2)

[mrickard]

Description

Customer reported an error when triggering an instrumented Lambda function via APIGatewayRequestAuthorizerEvent.
This PR makes the event matching more restrictive to avoid false positives.

How to Test

Added event fixture to test/unit/serverless/fixtures.js and assertions to test/unit/serverless/utils.test.js. Added unit test for API Gateway V2 websocket events in test/unit/serverless/api-gateway-v2-websocket.test.js

Related Issues

Closes #2857
Closes NR-354739

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.22%. Comparing base (763c0e6) to head (bcef1de).
Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2863      +/-   ##
==========================================
- Coverage   97.27%   97.22%   -0.06%     
==========================================
  Files         296      296              
  Lines       46625    46552      -73     
==========================================
- Hits        45354    45259      -95     
- Misses       1271     1293      +22     

Flag	Coverage Δ
integration-tests-cjs-18.x	73.17% <22.22%> (+0.01%)	⬆️
integration-tests-cjs-20.x	73.18% <22.22%> (+0.01%)	⬆️
integration-tests-cjs-22.x	73.21% <22.22%> (+0.02%)	⬆️
integration-tests-esm-18.x	49.88% <22.22%> (-0.04%)	⬇️
integration-tests-esm-20.x	49.89% <22.22%> (-0.04%)	⬇️
integration-tests-esm-22.x	49.94% <22.22%> (-0.04%)	⬇️
unit-tests-18.x	88.88% <100.00%> (-0.02%)	⬇️
unit-tests-20.x	88.88% <100.00%> (-0.02%)	⬇️
unit-tests-22.x	88.89% <100.00%> (-0.02%)	⬇️
versioned-tests-18.x	79.08% <22.22%> (-0.17%)	⬇️
versioned-tests-20.x	79.09% <22.22%> (-0.17%)	⬇️
versioned-tests-22.x	79.10% <22.22%> (-0.17%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[bizob2828]

This logic has been pared down quite a lot. Are we certain this won't regress on the driver for the initial refactor of this file?

[mrickard]

@bizob2828 That's a valid concern. The initial driver was for support for API Gateway V2 events, which are structured differently from V1, particularly regarding the path, httpMethod, and headers (or, in the case of v1 multiValueHeaders) we used for determining whether or not an event represented a web request.

Initially we focused on identifying v1 vs v2 events in this commit. While this takes a similar approach to the current PR, it checks event version first. That could result in non-web requests being matched as web requests, as non-web-requests can use API Gateway v1 and v2.

This subsequent fix was driven by a similar issue to the current issue: a non-web request was identified as a web request. The fix matched on more event properties, and better identified web requests because of that, but it was too restrictive, excluding ALB web events from being identified as web requests.

Subsequent fixes added explicit support for v1 REST events and for ALB events. These focused on those event sources, but did not focus explicitly on identifying web from non-web requests. (ALB matching is removed for this PR because its web-request signature is the same as that of API Gateway v1.)

One distinct advantage of all of these fixes is that they've added tests we can use to check whether or not the matching behaves as expected. Those are still passing for this refactoring.

[bizob2828]

one tiny suggestion on the code comments, but otherwise LGTM