IP SLA is one of solutions. It's easy to configure and pretty intuitive to use SNMP MIBs. IP SLA is based on synthetic packets sent periodically by your configured schedule. It will show how well these packets were sent (latency, errors, jitters and few more parameters). These parameters can be read from 22.214.171.124.126.96.36.199.42 SNMP branch.
Another thing is if you plan to monitor how your existing traffic is working on every sincle class: errors, drops, queues, etc. In this case you need to monitor that QoS's SNMP OIDs. Since you haven't sent your config lines so far, in general I can recommend you find your QoS which you use and browse through according MIB and see what parameters you can monitor and which OID they are assigned to. Check following link for MIBs you may need. At least you can find what parameters you can monitor and which OID has your required parameter. http://forums13.itrc.hp.com/service/forums/questionanswer.do?threadId=1399700&admit=109447627+1263488368517+28353475
Were you able to find a solution besides the IP SLA option? I'm curious to know how you were able to accomplish this as we are trying to measure the same thing in our environment ... NNMi 8.13 with Perf SPI and Cisco nodes.