I’m using the NUT integration to monitor a UPS on an external server that runs stand-alone NUT (i.e., not the NUT add-on that would run on the HA machine). The communication works fine but the polling interval seems to be hard-coded to 60 seconds with no apparent way to change it.
Years-old search results suggest an option in the integration config to reduce the polling interval down to 10s, but this seems to be outdated because the config (now) only contains the server address and credentials. The actual NUT server is configured with a 5s polling interval. Is there a way to reduce the polling interval of the integration in order to make it more usable, especially for alarms?
What trigger did you use for the polling automation?
How much time do you reckon is better to poll without increasing CPU usage too much?
Have you noticed disadvantages?
A 5s polling interval is plenty for a “responsive” display. I don’t see any CPU spikes or adverse effects, but I also have everything running on an i5 10th gen with 32GB RAM (Proxmox with NUT directly running on the host [] and HA in a VM), so that kind of “load” is background noise.
Also-also, the 5s polling only updates 6 important sensor values (online/offline status, voltage, load percentage, power, battery runtime, battery charge). NUT provides a bunch of other data (battery capacity, type and whatnot) which is not time-critical, so for that I have another automation which polls every 30 seconds.
] and HA in a VM), so that kind of “load” is background noise.