Open balis opened 8 years ago
How do we now a sensor is malfunctioning? Are malfunctioning sensors only those not reporting values for more than 24 hours? Recent analysis showed that there are sensors pushing bad values as well. Should we start with the ones not pushing values? What if a device aggregate contains only some malfunctioning sensors? Should we indicate these somehow as well?
I suppose it's easy to discover sensors which aren't reporting readings - this information is directly returned in each Parameter
object. Regarding out-of-range values, a simple threshold could be established in the UI as a stopgap measure (e.g. display sensor as "malfunctioning" if the reported temperature is above 50C or below -50C - from what I can see the "bad" Neosentio temperature sensors tend to report outlandish values of 400C or more - this is confirmed by the Neosentio guys who I met with today).
I remember one of the discussion during one of official ismop meeting, where this problem was mentioned. Daniel K. proposed to have valid parameter values range recorded in parameter metadata. Then is can be used to discover (maybe similar like non functioning devices - using delayed job) malfunctioning sensors on the DAP side and expose this information using REST API.
What do you think?
A. Borecka suggested that malfunctioning sensors should be somehow highlighted