Version

Argent Advanced Technology 3.1A-1410-A and earlier

Date

Friday, 19 Dec 2014

Summary

A Relator can use dynamic scheduling and option ‘Re-check Every x Minutes’

Click For Full Size

When A Rule is broken the first time, an Event is fired as expected

The Supervising Engine starts running a re-check every x minutes as specified in the Relator setting

When Argent AT runs recheck, it does not fire Event if Rule is still broken

A coding error in dynamic scheduling causes recheck is executed at the spot that a normal Relator execution is scheduled

As results, Supervising Engine might not fire new events for the same condition until recheck stops when the condition is corrected or maximum re-check time has been reached

The issue persists even the original Event is answered or purged

The issue does not happen for Relator using scheduling with fixed start time or do recheck using normal Relator check mechanism

The issue has been addressed in 3.1A-1501-A or later

Technical Background

After the first event is fired for the Relator, Supervising Engine will run recheck on the broken Rule every one or a few minutes as specified in the Relator

For the recheck, no event should be fired when Rule is found broken

The initial Event has already notified customer the condition

The recheck is mixed with the regular Relator execution

For example, a Relator is scheduled to run once every 10 minutes

When a Rule is broken, the Relator should do recheck every 1 minute

After the Rule is broken for the first time, Supervising Engine should run 9 rechecks then 1 normal execution and repeat the pattern every 10 minutes

If the Relator is dynamically scheduled, in Argent AT 3.1A-1410-A or earlier, Supervising Engine will run 10 rechecks every 10 minutes instead

The normal execution is wrongly replaced with recheck

As results, no new event will fire even the original event is answered or purged

The issue does not always happen even for dynamic scheduling

It happens if the Relator interval is multiple times of recheck interval

For example, if Relator interval is 60 minutes, the issue will happen if recheck interval is 30 minutes, but the issue will not happen if recheck interval is 7 minutes

The issue can be also ‘resolved’ if the condition is corrected then Rule is broken again

When the condition is corrected, the recheck will stop

When the Rule is broken again, Relator is executed normally, not a recheck

As results, a new Event will be fired

Most customers set up recheck for automatic resolving events

Customer does not expect new Event if condition persists

That’s probably why the issue has not been reported until recently

Resolution

Upgrade to Argent AT 3.1A-1501-A or later

For customer who cannot upgrade immediately, he can reconfigure the Relator to use option ‘Use Normal Relator Check Mechanism’, or use fixed start time in Relator scheduling