This isn’t a suggestion. TAC will ask for the “show status” output from the CLI. If any requirement is violated, your case is closed.
Cisco provides specific OVA files based on the "User Count" or "Node Size." You must choose the template that matches your scale: cisco cucm virtualization requirements
Most importantly: Set DRS to partially automated or manual for CUCM VMs. This isn’t a suggestion
Do not use thin provisioning for CUCM—ever. The performance penalty during write bursts can trigger node outages. Cisco provides specific OVA files based on the
Resource allocation for CUCM is not dynamic; it is static and deterministic. Unlike data workloads that benefit from "thin provisioning" and memory sharing (ballooning/swap), CUCM requires dedicated resources.
Cisco primarily supports for CUCM virtualization.
I once saw an engineer deploy 3 CUCM publishers on a single 8-core host, each with 4 vCPUs. The math looked fine (12 vCPUs on 8 cores? Not fine). Call quality degraded during peak hours. TAC refused support until the VMs were moved to dedicated hardware.