I am looking for some details on the PI Performance Counter named "Event Frames Written". This counter also shows up in the Analysis Service Statistics Dialog as "EventFramePublicationCount".
The description of the metric is stated as "Event Frames Written: Number of Event Frames published since service start"
Those appear to be the same metric as they show the same count.It shows a value of over 65,000 event frames published in just 3 days (restarted service 3 days ago). This seems extremely high for our normal operations. I a trying to track down the cause and also working with support on some analysis performance issues. I wanted to get some additional feedback.
In an effort track down what events are firing and contributing to the number, i looked at the event frames written in the last hour and calculated a total of 591 by looking at the performance metric value an hour ago as compared to now.
I then used PSE to search for all event frames that occurred in the last hour, to my surprise the search only returned 28 event frames. We only have 1 main AF database. There is a System Monitoring AF database, there were 0 EFs generated in that AF database in the last hour.
As an additional check , i ran this query against the PI OLEDB Enterprise connector and observed the same result, 28.
SELECT * FROM [MyDatabase].[EventFrame].[EventFrame] where starttime > '*-1h' order by name, starttime
28 versus 591 is a huge discrepancy. If anyone could shed some light on the true meaning of "Event Frames Written" that would be helpful