Flexible GC workload

[carpawell]

Is your feature request related to a problem? Please describe.

I'm always frustrated when we delete containers with millions objects and it may last for days.

Describe the solution you'd like

Be able to tune the number of objects being deleted based on a current load. If a node does nothing, it may try harder, but if we are being benchmarked, removing objects from disks may be stopped for some time.

Describe alternatives you've considered

-

[cthulhu-rider]

when the free disk is approaching zero, the collector should increase its power. At the same time, the increased load of incoming data should not trigger its sleep as proposed