Update multiverse jobs in CI to avoid resource issues in github actions

[tannalynn]

So, this does seem to work to prevent the github communication errors/running out of resources thing from happening, yay!

scheduled CI run: https://github.com/newrelic/newrelic-ruby-agent/actions/runs/12772900369
I saw a couple of random actual intermittent test failures, like a timeout or openai complaining about not having a token, BUT I haven't seen ANY github communication errors come up on there and I tried rerunning it several times to give it as many chances to fail as I could.

I don't love how much bigger the files are now and all the duplication I had to add.
But it's better to have an annoying file than have github actions run out of resources all the time though, so ¯\(ツ)/¯ oh well

Here is what changed for multiverse:

• multiverse: removed all services and runs most our multiverse like normal
• multiverse_services_1: elasticsearch, mongo, bunny
• multiverse_services_2: redis, sidekiq (bc sidekiq uses redis), memcache
• multiverse_services_kafka: rdkafka, ruby_kafka
• multiverse_services_mysql_pg: active_record, active_record_pg

I wanted mysql running with very few other things since that was mentioned in some of the comments we saw as being a problem.
I also didn't want too many services running in the same one just because I wanted to minimize the number of services each group was running to avoid resource issues.
Also I wanted kafka on its own because we've seen some weird issues with rdkafka in the past so just wanted those alone in case those things continue happening.

resolves #3004

[github-actions]

SimpleCov Report

	Coverage	Threshold
Line	93.54%	93%

[kaylareopelle]

I'm so glad you fixed this problem! 🎉 🧶

Seeing all the duplication, I wonder if we could use a composite action to reduce some of the steps we write again and again?

Here's the GHA doc on reducing duplication that taught me about this concept.

Rather than push things up into a separate repo, I imagine we could just use our actions directory.

I also understand if that refactor seems like too much for now. We could create a ticket and add it to the backlog if you prefer to leave things as-is.

[tannalynn]

I'm so glad you fixed this problem! 🎉 🧶

Seeing all the duplication, I wonder if we could use a composite action to reduce some of the steps we write again and again?

Here's the GHA doc on reducing duplication that taught me about this concept.

Rather than push things up into a separate repo, I imagine we could just use our actions directory.

I also understand if that refactor seems like too much for now. We could create a ticket and add it to the backlog if you prefer to leave things as-is.

I think looking into that would be a good idea, but I do think it would be better as a future improvement. It would be my preference to leave the duplication for now, just so things are working again at least. That way I can pick redshift back up again to make sure we don't fall behind there, since that's been tricky.
But i do like that idea, i definitely don't love the duplication! lol.

[kaylareopelle]

Sounds good! Let's create an issue after this PR is merged.