Improve CI capacity dashboards

Currently we have CI capacity dashboard: https://testresults.qt.io/grafana/d/DQHnzeK4z/ci-metrics-coin-capacity

However it doesn't have a view a for all the use cases:
For any given pool of machines (say macOS 11) and a given time span

• What percentage of the time that they have all been busy has there been other jobs waiting? (overloaded capacity)
• What percentage of time that some machine has been busy has there been other machines idle? (spare capacity)
• What job IDs have been seen in the time span given? (are we running jobs we shouldn't)
• How much time was spent, in percentage of total time spent in pool, of each job? (how heavy is each job)