Nutanix documentation on rebuild capacity reservation and resilient capacity emphasizes that some storage space must remain effectively protected so the cluster can recover from failures and maintain resiliency. If a cluster consumes all capacity beyond that protected threshold, it can reach a point where it no longer safely accepts writes because continuing to do so would violate resiliency guarantees. The symptoms in the question—stopped accepting writes and VM outages—align with a cluster hitting its resilient capacity limit, not merely an EC optimization limit. ( Nutanix )

This distinction matters. Erasure coding limits affect efficiency opportunities, but they do not by themselves explain the cluster refusing writes. A degraded container after a node failure may contribute to pressure, but the wording centers on cluster-wide behavior after rebuild-capacity reservation is enabled. Nutanix’s resiliency logic is designed to protect the cluster from entering an unsafe state, and one enforcement mechanism is to stop admitting additional writes when resilient capacity has been exhausted. Therefore, the most likely cause is C: the cluster has reached its resilient capacity limit. ( Nutanix )