Description

This PR introduces a descheduler component to automatically rebalance virtual machines (VMs) across cluster nodes. Functionality becomes active only when descheduler module is enabled

• Load-based rebalancing:

  • Triggers when node CPU utilization exceeds 80%
  • Uses LowNodeUtilization policy for balanced resource distribution
  • Ensures even workload distribution across cluster nodes

• Affinity/anti-affinity compliance:

  • Handles cases where nodes no longer meet requiredDuringSchedulingIgnoredDuringExecution rules
  • Automatically migrates VMs to compliant nodes when placement constraints are violated

Why do we need it, and what problem does it solve?

Without active rebalancing, uneven VM distribution can lead to node overutilization (e.g., CPU >80%), causing performance degradation, latency spikes, and potential node failures.

Changes in node availability, labels, or taints may invalidate existing VM placements that rely on requiredDuringSchedulingIgnoredDuringExecution rules.

What is the expected result?

Checklist

• The code is covered by unit tests.
• e2e tests passed.
• Documentation updated according to the changes.
• Changes were tested in the Kubernetes cluster manually.

Changelog entries

section: core
type: feature
summary: Add automatic rebalancing of virtual machines to optimize load distribution among cluster nodes based on CPU usage threshold (80%) and affinity/anti-affinity rules. This functionality is activated only when the `descheduler` module is enabled.