If sensors are continuously increasing values, then the events in a period should be merged a bit differently.
Check what the current design does to these events in terms of computing the period value, and change it accordingly.
We should use the last value reported in the previous period for the delta ?
Closing this one. The current thinking (#97) is that we should send batches of events to the daemon, and let the control loop do the analysis it wants.
If sensors are continuously increasing values, then the events in a period should be merged a bit differently. Check what the current design does to these events in terms of computing the period value, and change it accordingly.
We should use the last value reported in the previous period for the delta ?