Good morning, i have a strange problem with a notification prematurely closing.
I'll try to figure out the environment as best as possibile. I have a notification with multiple conditions, one of these is the one which gives me the problem and it is:
Value is Good and PrevEvent('Value', '*') <= '*-30m'
and this notification checks periodically every 2seconds, notify only on change in status (email), with a resend interval of 1 day and non repetition interval set to 0.
The tag is an opc tag, with exception set to 0 Dev and 10min and a compression set to 0,03Dev and 8hours.
and Values are good during the entire testing period
now when the last event in the archive is 30minutes old for this tag, a notification mail of "frozen" is sent correctly reporting inside the triggering timestamp. But immediately after that a closure notification is sent reporting that the condition is not anymore true ("normal" or "not frozen") and the notification is closed with a reported triggering timestamp of only 2 seconds after the initial trigger (which is "casually?" the periodic cycle of notification checks). BUT the tag is always at 0 without any variation in archive...so it's a wrong notification trigger. (just one time happened that the triggering timestamp of closure was 4 seconds after the initial one)
After several tests i found that if i put the compression to off, and set the compression settings to the exception instead (so Excep Dev 0,03 and 8 hours - Compression OFF), everything triggers correctly as expected, and no more fake closure happens.
So My question is: Is there a problem/bug with notification and compression related archive events during the evalution of conditions? Am i missing something? Anyone have had the same problem?
Thanks a lot for any advice
Additional Environment Informations:
PI Server v3.4.390.16
PI AF Server 2016 v126.96.36.19944
PI Notification 2012 v1.2.1205.10