There are really only a few things that can affect performance of NetApps - CPU load, and disk load.
Assuming you want to track this info over time, and graphically, and not just use CLI tools:
CPU load is available by snmp. The consistency point statistics (also available by snmp) are secondary indicators of CPU load and write activity.
Cache age isn't very actionable by itself - it can be a starting point to see if your workload would benefit from a PAM card.
Disk load is an issue as it affects latency - so what you really want to track is latency of read/write requests per volume. And ideally, have information about which volumes are on the same aggregate, as they are the ones that can impact each other.
So if there is a performance issue caused by read or write request latency, you want to see which volumes on that aggregate had unusual operation load. Or track disk utilization for that aggregate, to see if you need to add more spindles, or migrate volumes.
Unfortunately, volume performance info is not available by snmp. You can use the SDK to get this information (http://communities.netapp.com/docs/DOC-1110), and there are various scripts floating around that can help.
CPU is .1.3.6.1.4.1.789.1.2.2.2.0
Also useful to track is operations per protocol (such as NFS Ops - .1.3.6.1.4.1.789.1.2.2.6.0; CIFs ops - .1.3.6.1.4.1.789.1.2.2.8.0)
But the more useful volume latency, operations, throughput, etc are not there in snmp.
If you dont want to build/manage your own monitoring system, check out LogicMonitor.com which automates all the above monitoring with no configuration required..
