We have found that the 'Robot Inactive' alert is great for letting our Server Team know that monitoring is not working properly and in most cases that the server is down or not responding. However, they would like a definitive message\alert that the server itself is completely unresponsive. In some cases the Robot Inactive doesn't show up because the agent is still communicating with the hub but a user or app cannot access the filesystem (server process usually failing).
We do have another app running that attempts a dynamic mapping to a UNC path and if it fails (because it can't see the filesystem) it will generate a failure message. They have stated that this process has proven to be 100% accurate in letting them know a server has indeed failed or not responding.
Any ideas how to integrate something of this nature? DIRSCAN maybe?