I'm afraid we are not collecting this metric at the moment. Could you tell us more about your use case? Do you want a metric that will show how much time a host was up for the specified timerame (i.e. 99,998%) or the actual uptime metric? It's different because host could be restarted because of some planned maintenance, uptime counter will be reset in this case, but it doesn't mean that the host has bad "uptime" KPI. So uptime counter could be misleading in detecting how good was uptime KPI for a host and then anyway you have to summarize all values between the resets, but what about gaps in monitoring? So for me it looks difficult to calculate a correct uptime using the metric, but maybe more research is needed, please let me know your way of using the metric - I would really apprecaite it.
Monitor uptime in SCOM reporting on the other hand provides a precise estimation for how long a certain aspect of host availability was in downtime. I think if you select for example "Veeam VMware: Host connection and power state Alarm" in the report above it will tell you precicely for how long a host was unavailable from vCenter perspective, not ideal of course, but without a vCenter, host can simply run a VM, many virtualization services provided by a vcenter will be unavailable (such as vMotion, DRS, storageDRS, HA, etc.). As a key factor - monitoring of the host is going to be unavailable, so even if we collected uptime metric, we won't be able to grab it if vcenter doesn't "see" a host.
I would really appreciate more information about your use case.