Technically there's no limit on the cluster per se. You can have 255 snaps per volume & the # of volumes per system (which I think is 500 per controller, but I'd have to look it up).
The NDU guidelines exist due to failover times. This is particularly important if you have LUNs on your system. If it's NAS-only you may be able to get away with it anyway unless you have an app (like Oracle RAC) which has similar 120 second requirements.
I believe the snap NDU limit is 10 * # of volumes supported for NDU which varies by platform, but is typically between 100-300.
I think Upgrade Advisor should tell you what the max # of snapshots for your platform when it says you are over the limit (or it did when I maintained the code )