Is this device configured to have it's availability and response times calculated via SNMP? If so, check if you can get to the device from the managing polling engine over SNMP (snmpwalk.exe, found within the SolarWinds install directory, will allow you to do this better than a simple TELNET test on port 161).
I expect that if the device is up and and pingable from the polling engine, then you'll need to list resources on the node and swap availability and response time back to ICMP to cure the node down problem you're having.