I had the same thing happen about a year ago with a customer on AOS 8.2, MM, two L2 clustered 7220's and a few hundred APs. TAC ultimately had me upgrade the code to a slightly more recent version, which made the memory issue go away, and I haven't heard of the problem resurfacing yet.
While this fixed the issue for me, I was unclear as to what may have caused it and if there was anything I could have done with the configuration or by restarting processes in order to fix it.
If you hear back from TAC as to how to troubleshoot and fix this outside of code upgrades - or what the root cause is - I'd love to hear.