There are a lot of measurement in every node of the NSS including the MSC-server, MGW, HLR, EIR and each has a more than hunderds of measurement. When you are saying KPI it means the Key performance indicators i.e the most important ones. Gos of the routes, SS7 signalling load and the access and the failure. From the Access KPI you will get ASR and the ratio of each failure like unknown failure for a call, failure due to caller busy and all this things, the monitoring of the access ratios will lead you to find out problems if they exist and you will use the other measurement to find out the failures like if the SCP or IN has a problem.
That is how I do I, any new ideas or recommendations