We have various interface nodes that buffer to multiple logical servers, one of which is a collective. I'd like to set up a display and notification rules to check for queue corruption for the collective members. I'm assuming a client who's connected to the primary member won't check the secondary's perfmon tag data if the primary buffer queue gets corrupted and its tags go stale. Here are a few ideas I have:
1. Add Perfmon tags for all physical servers to each buffered server. I can then use the Perfmon tags from a standalone server to monitor the collective member buffer queues. This may not be feasible in my case because the standalone server may not be acceptable or available to use for production monitoring.
2. Set up a remote Perfmon interface. In my case, there could be constraints in altering the existing environment.
3. Use ping data to deduce that buffer is failing. If ping tags are good but perfmon fails it would suggest a general interface/buffer issue, but ideally I'd like to know if a queue file is corrupted and which collective member it belongs to.
4. Custom AF SDK app. Not sure where to start with this.
I'm hoping to figure out if these options make sense technically, and if I'm missing a better approach to this.