OpenTelemetry Cucumber Instrumentation

[Ugzuzg]

I'm working on an OpenTelemetry Instrumentation for Cucumber.

🤔 What's the problem you've observed?

• The documentation mentions that deep requires capability will be removed in a future release: https://github.com/cucumber/cucumber-js/blob/126772d66d0889d08421a30bf6e79fa9ec95e52b/UPGRADING.md#deep-requires. I'm relying on a deep require of a TestCaseRunner to create spans for scenarios. Creating an issue just to have a discussion and make sure that you're aware of this use-case.
• Scenario retries are done in a way that makes it impossible to group each attempt under one span.

✨ Do you have a proposal for making it better?

• Would you be open to a PR that extracts the code that attempts to run a scenario into a separate method of TestCaseRunner, so that it is possible to create a span for each attempt?

📚 Any additional context?

Passing scenario:

A trace of a scenario that was retried (current implementation):

[davidjgoss]

First off thanks for raising @Ugzuzg and explaining your use case, I’m a fan of otel so this is really great to see.

To your question: yes more than happy to take a PR for that refactor you describe.

But also: it would be much better if you didn’t have to reach into internal functions to do the instrumentation. We’ve been working on a plugin system for extending Cucumber and this seems like a good problem to help move that along. I’ll study your otel PR a bit more for specifics but essentially I’m thinking of having granular lifecycle hooks for injecting custom logic before and after invocations of test cases, steps etc.

[Ugzuzg]

@davidjgoss, the instrumentation was released 11 days ago.