[Reactor] Bug when sending MQTT boolean payloads

[3rdStng]

Super excited to see in the upcoming release notes about MSR being able to poll HE z-wave devices. I've attempted to use the native HE poller and in less than an hour after installing it, I was removing it. I only have a small number of devices that give me trouble, but I wanted to verify that I'm pre-building my YAML file correctly.

poll_frequency: 60000  # poll a device at most every 60000ms (60 seconds)
poll_entities:
     - id: 110  # Hallway Light
       interval: 300  # 300 seconds = 5 minutes
     - id: 111  # Kitchen Light
       interval: 900  # 900 seconds = 15 minutes
     - id: 112  # Disco Light
     - id: 113  # Ballroom Light

Am I correct in my thoughts? Based on the above snippet, the default polling will be every minute for the Disco and Ballroom lights? But because the Kitchen and Hallway light have a defined interval, they will be polled at those intervals instead? Or does there need to be an "interval:" definition following every id:?

[toggledbits]

The Disco Light and Ballroom Light will be polled at the default interval of 10 minutes (600 seconds). You only need to specify an interval if you want something other than this default.

The poll_frequency is not a default for the polling interval. The poll_frequency instead is the minimum time between device polls that the poller is allowed, to limit its ability to saturate your Z-Wave mesh. At the default frequency of 60000 (milliseconds = 60 seconds = 1 minute), one device per minute will be polled, regardless of how many devices are due to be polled. In effect, the interval becomes the minimum time between polls of a single device, and the actual interval depends on how busy the poller becomes. If you have a large number of devices being polled, or are polled frequently, then it's possible that polling will, at least for some devices, become chronically late. The next device to be polled is always the tardiest (i.e. the one that has been waiting longest), so every one will get polled.

[3rdStng]

Thank you. So if I have 10 devices, all with the interval set at 60 seconds, essentially it will poll one device at minute 1, the second device at minute 2.
In other words, one device polled per minute, or based on the poll frequency.

If I'm understanding this correctly, I'm better off upgrading my z-wave to z-wave plus devices and only poll those that are critical, and not upgraded, yet.

[toggledbits]

@3rdstng I think that's right. If an automation relies on the crisp response of a device that doesn't have it, polling is a poor solution and the device should be replaced.

A counter-example from my own home, though... I have a few old switches and dimmers controlling a guest bathroom in our basement, policing that neither the fan nor light are left on unsupervised for hours at a time. It's sufficient (in my view) to poll every 10 minutes, and the time delay of the auto-off automation is OK being delayed that much. As long as the lights aren't on for hours (or days) unoccupied, I'm satisfied.

[3rdStng]

@toggledbits said in Upcoming HubitatController z-wave polling feature:

I think that's right

It is correct. My polling config left the default frequency at 60000ms (1 minute) and I did not include the interval settings on my devices, of which I am polling 3 until I can get around to replacing them. Upon reactor startup, the first device in the list is polled, one minute later the second device, another minute later, the third device. Seven minutes from there, or ten minutes from the initial poll, device 1 is polled again.

Thank you @toggledbits for adding this feature.