On my Win2008 R2 clusters there is a performance counter called Cluster Multicast Request-Response Messages\Messages Outstanding.
Solarwinds includes it as a monitor on their Failover Clustering application template. One of my Hyper-V nodes keeps having this alarm go off (value greater than 0). I'm willing to remove this counter from the monitor but want to understand what it means first.
There is a Failover Clustering Performance Counters blog post on MSDN that describes the different Performance counters that are available.
In Perfmon on target cluster node you should be able to see exactly which resource is failing. If your environment is working fine with these occasional errors you can disable thresholds for this component monitor or increase current critical threshold (greater than 3).