An administrator enables NDB High Availability.

In an NDB high-availability deployment, multiple API server and repository VMs protect the availability of the management plane. If the primary API server becomes unavailable, NDB automatically fails over API responsibilities to an available standby server. HAProxy directs management requests to the functioning service, allowing administrators and automation clients to continue accessing NDB. The managed databases operate independently of the NDB API service for normal client connections, so an API-server failure does not automatically stop database workloads. Backup or patching workflows are not categorically canceled or suspended merely because one API service fails; their processing continues through the surviving HA components, subject to the state of the individual operation. Nutanix’s official highly available NDB architecture confirms that NDB automatically fails over to a standby VM when the primary API server or repository VM becomes unavailable.