Deployment autoscaler enhancements

[ekazachkova]

relates to #2639

This PR brings the following changes:

• a new trigger that provides ability to scale up node/replicas if pods utilisation is greater than config provided threshold (config.trigger.pods_utilization)
• a new rule that implements scaling down for transient nodes with no not-terminated target pods. The scale down process shall not applied if:
  • current cluster size is <= minimum cluster size (config.limit.min_nodes_number)
  • interval between last scaling process is nor greater than minimum acceptable (config.limit.min_scale_interval)
  • node is forbidden (configuration.target.forbidden_nodes)

To manage node scale strategy (SKIP/STOP) config.rules.on_not_running_target_pods can be used (Default: SKIP)