Hello, I am trying to find out how to write a where clause to filter out data older than 5 minutes for a Prognosis threshold I am building. Normally, I would not have to define a time range when I put together a threshold, but if I do not define a time window, each interval check that the threshold performs, immediately sends alerts even though the reference data is more or less an event log. AVHISTORY checks against the active log but does not behave like a HISTORICAL record (one that has the 'deliver differences' behavior) so each interval that passes I get another volume of emails referencing the same alerts as before. I have tried filtering using the TIME record but prognosis cannot resolve the type of this literal: TIME < 10 Minutes AND ACTION = "cha" AND OBJECT = "station" AND QUALFIER = "8452". My goal is to be able to instantly report critical changes to my PBX's (trunks, routes, signaling groups) via email using the dispatch manager. If there is another way I could accomplish this more effectively, that would be great.
Solved! Go to Solution.
Hi @Edmanuel_Ferrer,
TIME is actually a string field, but there is also a TIMESTMP field of timestamp type that you can use. So if you want to only look at data within last 10 minutes, you can apply the following filter:
TIMESTMP > CurrentTime - 10 Minutes
Also, if you check the 'Off Event' checkbox under 'Message Defaults' tab, then a condition that is true and remains true would only generate 1 alert, keep it open untill its false and then marks it closed. Without using the 'Off Event' checkbox, it will keep opening new alerts and automatically mark them as closed.
Hi @Edmanuel_Ferrer,
TIME is actually a string field, but there is also a TIMESTMP field of timestamp type that you can use. So if you want to only look at data within last 10 minutes, you can apply the following filter:
TIMESTMP > CurrentTime - 10 Minutes
Also, if you check the 'Off Event' checkbox under 'Message Defaults' tab, then a condition that is true and remains true would only generate 1 alert, keep it open untill its false and then marks it closed. Without using the 'Off Event' checkbox, it will keep opening new alerts and automatically mark them as closed.
Thank you for the swift reply Shoaib. I applied the recommended where clause string with no success. Upon further investigation, I found that Prognosis is not converting the issue time into the prognosis format, shown here:
Is there record I could reference or should I open a support ticket to address the blank TIMESTMP field?
Hi @Edmanuel_Ferrer,
you are right, looks like that field is not being populated. As you suggested, pls open a support case for further inverstigation.
Did the 'Off Event' not do the trick? You can also play with 'Log every' and 'Do not log' options as well.
Hi Shoaib,
I havent tried the Off Event since the AVHISTORY record is more or less logging events rather than changing states. I am, however interested in the 'Log every' and 'Do not log' options, unless they reference the log frequency section. I have tried setting the 'Log at most every...' to 5 minutes or so and my check interval is set for 1 minute. Once I started the threshold, I recieved seven emails per minute with the same batch of alerts coming in as if the 'log at most...' setting never took effect. Thank you for your time Shoaib, it is deeply appreciated.
That was what I was reffering to. One thing I'd mention is that if a new event is triggered, even for same issue, e.g username/password mismatch, it will be considered a new etry (based on unique identifiers), so in this case 'Log at most .' would not help.
Members | Likes |
---|---|
43 | |
36 | |
23 | |
17 | |
11 |