Multiple timestamps appear wrong/dated

[tjmccull]

Hi,

Been using this code for just over a week now and it has been working well! Thanks! That said I am observing that for multiple data points that the timestamp recorded appears wrong (to what I would expect) and in many cases very dated (ie over a year old in some cases). I suspect this may be an issue of the underlying driver but I am not sure. Disclaimer here is that I am brand new to these Edge drivers.

I presume your code is requesting (and logging) the timestamp as reported from the driver of the most recent state change reported, versus recording the timestamp of the monitoring poll.

Once data is in InfluxDB I leverage Grafana to create a custom dashboard to display the current state/trends of my sensors data. With these really old timestamps I need to set my query to go back hundreds of days which makes it really difficult to observe recent changes in the trends/data.

I have many Samsung Smartthings branded sensors (such as door, leak, motion). For the sake of this thread I will focus on the water leak sensors. In some cases the battery field timestamp is over 6 months old. I would have expected the battery to report more discrete changes in levels over time. Or the water state (wet/dry) is over 2 years old in several cases (to be fair they likely haven't detected water since I tested them at install point. Current driver is "Zigbee Water leak sensor" from SmartThings version 2023-01-24T21:01:14.51179409.

I guess my question are:

1. Are my assumptions correct that the timestamps are linked to that last state change as reported from the driver
2. Is there a way to log (into InfluxDB) the timestamp of the current poll (which would be the current state), and by polling every min I could gather/log the trend data.

Thanks

[tjmccull]

3. Is there a "refresh" API call that would update the timestamp?

[eargollo]

The system basically calls SmartThings and logs the last datapoint that SmartThings provides for each device, given that this datapoint is new.

I suspect that you are right on your assumptions. For some sensors, the last datapoint is of the last time the sensor was activated. For instance, I have an external thermometer that does a read every hour. The system will them log only 1 entry per hour even as it calls the sensor every 5 minutes but receives the same date of read. Also, the system records the "sensor date" as it is the read date.

It is possible to either also record the read date and time, or to add some configuration for it to replace the sensor date by the read date and, in this case, record the value at every read independently if it is a new value or just the same data from the past.

For that I would need to change the code to add a configuration per device indicating that for that device we want the system to log the read datetime instead of the "sensor date".

Would that do the trick?

[tjmccull]

Appreciate the response! Yes, having the ability to have a configurable setting that would allow the system to log the value at every read and record the read datetime (vs sensor date) would solve the challenge I am trying to solve! Adding that feature would be greatly appreciated. There are likely multiple levels one could consider having this configurable option set at; such as globally, by sensor type (under the monitor: portion of the config file) or at each physical device level as you suggested. I wouldn't think a user would need the flexibility to configure this at multiple levels, so picking one that is best for you to implement would be fine. Thanks!

[eargollo]

Just to give an update that I am working on it. I took an initial detour on making the code more unit testable first.

[eargollo]

Take a look at the latest release. You can add the option to use wall clock to a capability. This way it will update influx every time SmartThings is called and use the current time instead of the sensor time.

Let me know if that works.