OpenEverest now supports tolerations, node selectors, and topology spread constraints for scheduling

Instance scheduling in OpenEverest now supports tolerations, node selectors, topology spread constraints, and custom scheduler names, all grouped under a unified schedulingPolicy block.

Previously, ComponentSpec only exposed affinity, which was never the whole placement story. A component pinned to a tainted node pool needed tolerations, and zone spreading normally requires topologySpreadConstraints — both had to be handled outside OpenEverest or not at all.

The new schedulingPolicy struct lives in the common API because placement is not instance-specific. You can now set schedulerName, nodeSelector, affinity, tolerations, and topologySpreadConstraints in one place, and every component that supports scheduling honours the same fields consistently.

This is available in OpenEverest v2.0.0 Developer Preview 3. Existing instances using the old affinity field will need to migrate to the new structure before the next preview release.