When Reserve Rebuild Capacity is enabled, Nutanix reserves an amount of free space equal to the largest node in the cluster so that, after a node failure, there is enough space to rebuild full data resiliency. This effectively caps the amount of data that can be written to the "Resilient Capacity" (total cluster capacity minus the space of one node for RF2, or two nodes for RF3). Once the cluster's stored data reaches this Resilient Capacity limit, the storage layer (Stargate) stops accepting new write requests to guarantee that the cluster can still tolerate and rebuild from a node failure. Prior VM outages (from node failures) would have consumed part of the reserved rebuild space, pushing the cluster to this limit sooner. This behavior is specific to the Resilient Capacity threshold enforced by Reserve Rebuild Capacity, not to EC-X capacity limits or to a currently degraded 1N/1D or 2N/2D container fault-tolerance state.
Community Discussion