Morning,
I am looking for some general input with regard to a fairly complicated RTA.
I have some 1,500 devices reporting location and device health every 30 seconds, as well as on specific events, so sometimes much more often between the periodic reports. The issue I have is that Time to First Lock can vary, and it's feasible to see numerous reports at different times of the day with a null Lat/Lon/Alt/GPS Time in the transaction.
So, I have removed the geometry calculation from the Feed and do it in a calculation node inside the RTA if LAT is not null. The enriched schema is then carried with the transaction through a couple of different logic streams, one of which does an Incident Detect and then a dynamic geoFence.
Now my issue, if I get a PwrDwn event from a device (end of shift), there is no location data, but one of the close conditions for the Incident is PwrDwn, which is currently never seen as the LAT filter eliminated the transaction.
So my question is: is there any reference documentation on which processing nodes can be in a logic stream even if there is no geoLocation in the schema, and which will throw an error, crash the RTA, and result in potential data loss?
Note the new Choice (Route by expression) makes the initial logic much cleaner and easier to implement. So it's which nodes have a dependency on geometry that I need to understand.
TIA --
