This is a known issue affecting 6.7 U3 and has been addressed in vCenter 7.0.
Workaround:
To workaround this issue, vSAN Health service on the vCenter Server can be restarted by running following command:
service-control --restart vsan-health
There is no impact on any vSAN cluster expected.
If the alarm comes back, try the following steps:
1. Go to vCenter -> Configure -> Alarm Definition
2. Find alarm "vSAN online health alarm 'Customer advisory for <alarm name mentioned after cloudhealth>'"
3. Delete this alarm.
4. Restart vSAN Health service as outlined above.
The deleted alarm will be added back. This is due to we have a mechanism to check and re-register all the system alarms when vSAN Health Service restarts.