Service map phase II

For the next version of service maps, here are a couple of additional use cases we want to cover:

- Show health indicators on service maps. If a service triggered an alert, surface the status of the service based on the triggered alert on the maps so I can determine what services are not healthy.
- If machine learning is configured for the service, surface the anomaly score indicator on the map to enable a user to detect an issue prior to it having a widespread impact(show more than just the critical anomaly scores)
- Allow a customer to switch between what health indicator to surface(default should be alert based)
- If no alerts or ML job is set up, prompt the user to set those up.

Related design issues:
- https://github.com/elastic/apm/issues/222
- https://github.com/elastic/apm/issues/228

In addition here is a list of other UI modifications that we will address in this phase:

- Add info to connections between services https://github.com/elastic/kibana/issues/52870
- address loading state on service popovers https://github.com/elastic/kibana/issues/59342
- handle map resizing on banner interactions https://github.com/elastic/kibana/issues/56412
- support fullscreen mode https://github.com/elastic/kibana/issues/56531