docs(threat-hunting): add ADR 0002 on usage of slow accessibility selectors in Jest RTL tests

[kapral18]

This PR introduces a new ADR concerning the usage of heavy react-testing-library selectors like ByRole, ByLabelText and ByText.

This ADR is images-heavy and also contains a separate tests benchmarking repo https://github.com/kapral18/rtl-selectors-perf-test to further expand on the findings in the "problem and context" section.

The decision will be finalized after team/cross-team discussion and before merging this PR

[elasticmachine]

Pinging @elastic/security-threat-hunting (Team:Threat Hunting)

[elasticmachine]

💚 Build Succeeded

• Buildkite Build
• Commit: 2cbc208

Metrics [docs]

✅ unchanged

History

• 💚 Build #252324 succeeded 96e2f72
• 💚 Build #251732 succeeded b3c3e72
• 💔 Build #251728 failed 111e312

cc @kapral18

[elasticmachine]

🤖 Jobs for this PR can be triggered through checkboxes. 🚧

ℹ️ To trigger the CI, please tick the checkbox below 👇

• Click to trigger kibana-pull-request for this PR!
• Click to trigger kibana-deploy-project-from-pr for this PR!

[PhilippeOberti]

This is an amazing write up, thank you!

A few thoughts:

• while option 3 would ensure that tests are running extremely fast, it would - like you mentioned - go against RTL best practices, which is not good.
• option 1 seems to me that we would basically not changed anything to what we're doing today. It is a viable option. Tickets are automatically created when tests take more than 5s to load/run, so these tests could be investigated individually and we could work on improving only the offenders...
• option 2 seems to me the best but like you said difficult to enforce...

I know it's not the answer you'd like to hear, but it seems to me that we could not change anything to what we're doing right now, and just focus on improving the tests that do take more than 5 seconds to run today...
I don't see option #3 being enforced, and I also don't see another good alternative...