[Logs+] Enforce dataset names

[Kerry350]

Summary

With the introduction of #160777 we are able to create basic custom integrations.

Currently the dataset name is always prefixed with the integration name. We want to adjust this logic so that the dataset name always adheres to some rules.

Ideally #163788 should be implemented first, so that all of this logic is hidden away in the package.

ACs

• Generating the entries no longer hardcodes the integration name prefix
• The custom integrations API has server side validation for the dataset name that allows two options:
  • A dataset name that is exactly the same as the integration name (e.g. given an integration name of server_logs, the dataset name could be server_logs).
  • A dataset name that is the integration name and a period plus a custom dataset suffix (e.g. given an integration name of server_logs and a dataset suffix of errors, server_logs.errors would be valid).
• When a custom suffix is not supplied we default automatically to matching the integration name exactly (as in, callers of the API will default to this).
• Locations that call the API (currently only the Observability Onboarding plugin) ensure that the user is always aware of what the "final" dataset will be. E.g. in the UI we will tell the user that the integration name will be used automatically, or that their suffix will be prefixed with the integration name. There should be no surprises with the prefix.
• Should also be backed by client side validation to save a trip to the server.

[elasticmachine]

Pinging @elastic/infra-monitoring-ui (Team:Infra Monitoring UI)