I'd be surprised if it was a server performance issue. In our experience analyses whilst being memory hungry, are pretty efficient, we have tens of thousands of them running quite happily.
Are you able to post the analysis setup and the target tag config (compression settings etc)? That might give us some clues.
Is this running in Prod? Do you have a dev/test system to compare it with?
7 of 7 people found this helpful
It looks like a data compression issue because the back-filled data is uncompressed (all values written) since it is out of order and the real-time data is compressed. Does it really lag, or is it the data density of real time is less dense than that of the back-filled data? The data is relatively linear for many periods so I would expect no values during those times if you have compression on.
I agree with Dan Fishman here. I suspect that if you tracked the snapshot of the blue tag while the analysis is running that you may see it updating every 5 minutes (unless your the exception settings except it out). However, after the fact, only the events that meet the tag's compression settings are archived.
Please let us know if taking a look at these settings helps!
1 of 1 people found this helpful
There are quite a few performance counters for PI Analysis Service. You can take a look at those to see if you're lagging behind. I suspect you're not. Based on what I'm seeing on your screen shot, I would tend to agree with Dan that PI Analysis Service is calculating every 5 minutes, but the calculated outputs are being compressed out due to the compression deviation setting of your analysis output PI Points.