AM trying to morph DCRUM to work with our alerting ticket system for the DCRUM heartbeat.
Heartbeat is sent every 10 minutes, If an event is not seen for 30 minutes, then an alert ticket is fired from external system to notify the possibility of DCRUM not available or working properly.
During the nightly tasks, DCRUM stops sending events for about 2:30 hours, until all the tasks are finished. Therefore no event is sent to external alert system and an alert is fired stating the heartbeat is missing.
If the DCRUM event is closed prior to the tasks running, the external system will not be expecting the heartbeat.
I need to send a close event at a specific time, which DCRUM can't do. Also can not close heartbeat event as it is based on a metric that always kicks event.
The close event just has to have the message needed for the external app to think it is closed. Does not have to be the original DCRUM event changing.
If I can get an event to fire at a specific time, I can work around the external systems limitations.
Then the DCRUM/NAM scheduling tools will not help you either. The scheduled tasks are based on data time as well, and so will also be subject to the same pause from the nightly tasks.
I would recommend setting a maintenance window so the heartbeat is not expected during the nightly tasks, or switch monitoring to an API query that confirms the report server is up, running, and either processing data as expected or processing the nightly tasks.
Erik, let me add one short comment to this statement:
There are two kinds of scheduled tasks in NAM, first is as you said based one data samples time, with all the consequences you mentioned above, but there is also 2nd kind of tasks, with timeline independent from samples time, based on real time.
We're working on solution that will fit best for Randy