Extracts the SQL challenge into its own track to make it easier to have separate dashboards for the SQL tasks and to reduce coupling between NOAA and SQL benchmarks.

This PR also addresses some other issues that have been identified in the meantime:

- Force-merges the index into a single segment to reduce variability between runs
- Adds the `wait-until-merges-finish` task to ensure merging finished before running the queries
- Introduces a `query_percentage` track parameter as a convenience to get faster results when doing experimentations (to be taken with a huge grain of salt of course)
- Ensures that `page_timeout` and `request_timeout` are configured correctly such that the track can be run on revisions with or without elastic/elasticsearch#79360.

Apart from these changes, the queries themselves have not been touched.