Next time it locks, can you check some onther data (if you didn't already)?
1. The CPU usage for the VM is high?
2. On the resource allocation tab for this VM, is it ballooning or swapping (like a lot)?
3. Can you try to restart the management agents for the ESX server where the VM resides and check if with this you can access it from the console?
4. Does ping looses any package?
The RDP test you did was directly from an RDP client to the VM? It works in normal conditions?