My favorite approach to monitoring my PI interfaces is to just pick a PI tag from the interface that is very active and compare the last time the value was archived. I do this in a PE and use the prevEvent function so I am reading from the archive and not the snapshot. This is extremely important because we have had issues with PIBUFFSS which has caused havoc and I believe this approach will catch this issue. We use Pi Notifications to notify when the Pi Tag has not archived a value after a certain amount of time.
My problem now is how to best monitor the nodes in the collective separately. We recently had an issue where data was being archived to the primary but being buffered at the interface for the secondary. This went undetected for sometime because of HA masking and almost shut us down because the interface machine almost ran out of disk space.
So in AF, I wanted to have attributes that specifically would only get data from a defined node which I do not believe to be possible. Now I believe that I may need to have a standalone app that specifically queries for the data from a particular node and writes the result back to the server. This is a little complicated and I would like to avoid this.
Are there any suggestions to this architecture. Your thoughts would be greatly appreciated.