I'd agree with @Adduxi that the notable bump in minimum latency is not of itself important, but indicates that VM have made some change to network arrangements, and possibly that's network resegmenting to address capacity issues and behind the improvement. Despite that there's a packet loss incident lasting five minutes or so at 9:50am and it looks like a fair few spikes remaining of around 70-120ms, occurring roughly every half hour when the network's busy. Depending on the time of last restart you may have too many T3 timeouts on the upstream and that might explain the latency spikes.
If those T3's continue to increase at (say) more than one every 2-3 days there's a likely upstream problem. Keep in mind that any restart will zero the timeout counters. It also clears the downstream error counters, but as they don't seem to work reliably on the Hub 4 that's academic.
