Add support for separate resource pools with independent queue priorities.

[jankaspar]

• Each executer can specify resource pool it belongs to in the configuration
• Priorities are newly tracked per pool identifiers, in the beginning of scheduling priorities corresponding to the pool are loaded.
• Metrics are also produced broken down per pool. For queued resources metrics we count the job under each pool it could run in based on its requirements.

[ingvarm-gr]

If I am reading this loop body correctly, could it be written as:

  poolReports, ok := result[report.Pool]
  if !ok { 
    poolReports = map[string]*api.ClusterSchedulingInfoReport{}
    result[report.Pool] = poolReports
  }
  poolReports[id] + report

Not suggesting that it gets rewritten that way, primarily checking my understanding of the code.

[jankaspar]

yes, your version is equivalent and more idiomatic, will update it.