I'm curious how others out there monitor their virtualization infrastructure. Specifically, do you run your NMS as a VM inside your production environment, as a VM in a management / monitoring cluster, or as a physical server that is external to your virtualization altogether? I see pros and cons for each:
Doing something else? I'm interested in that, too.
I would choose number 2 above, it the best option I think. Or maybe create a separate "Monitoring cluster" that runs only monitoring tools and does not interfere with production and management.