Well the hosts probably were happy before, however in a cluster (HA enabled) there has to be sufficient overhead to pick up the VM's from the other host. So when you have HA, the resources are 'reserved' for those VM's. Try turning HA off to see if the alert goes away.
If you only have 2 hosts, 50% of your machines are reserved for failover. The formula is number of machines / HA reserved reservation. The more machines you have, the more resources you have left over for failover. If you don't need HA, you can simply turn it off, that will save you a lot, especially with just 2 machines. HA is basically critical, can't go down, and automatic startup, but there is glitch, the VM's have to fail (shutdown) before they can be picked up by the new host, so the VM's will still be down for a few minutes. So unless you have high critical need, you probably don't need HA.
If you click on the cluster, you should see a status, and if the alarm was triggered, you should see a message to the affect of "There isn't sufficent resources to satisfy HA for cluster". Otherwise it could be something else triggering the error ( like a VM or something that is over it's CPU load) and it highlights the cluster to notify you there was a problem. Usually it just one host with a yellow triangle, but since it's a red mark on the cluster, I thought it could be specifically related to cluster only errors, that's why I was suggesting HA.
