I recently migrated from Vera to Hubitat C7. I have 45 z-wave plus devices and 11 older zwave in the new network for a total of 56 devices. After initial setup and then letting the network settle for a few weeks I am occasionally experiencing lags and some odd behaviors. I have about a dozen Basic Rules to automate lights and a few Rule Machine rules to automate shades. After building the network and rules, I'm experiencing a few problems:
-- I have 7 Leviton plugs and switches that report 9.6 kbps even though they are z-wave plus. 3 are within 10' of the hub. (I have 4 Leviton that properly report 100kbs). Is there any process (repair or configure etc) to try to get the slow ones faster?
--I have zooz zen77 dimmer that falsely randomly reports that it is was turned "on 10%" from the switch. Can this be a configure problem or a bad switch?
--I built a Rule Machine rule to Refresh all the devices every 30 minutes. It doesn't seem that this would clog the network but is it possibly at fault for the lagging?
-- When the devices were initially added/configured, the default on most was to turn on debug logging and description logging and I left it in that state. Can this slow things down?
--I built a rule that turns All Off that runs on a schedule. I'm turning about 10 devices OFF then delay about 4 sec, then the next batch of 10 etc. Occasionally the rule doesn't turn everything off. Do I need more delays?
--I have a few devices that I added to the network but aren't plugged in. Will this slow things down? Should I remove them until needed?
--I added maker.api to interface with Hubivue and a geofence app . Also running Homekit and Homebridge for Apple Home dashboard. Can the hub handle all of this or would it be better to remove one of the dashboards?
When you have good & well-placed repeaters and no ghosts in your mesh, then it's best to just let the mesh sort itself out instead of trying to bend its will -- that can take several days.
Debug logging isn't necessary (or desirable) to leave on unless troubleshooting an issue. IIRC, many (all?) devices have it on temporarily (30 mins or so?) to help verify stuff when adding it as a new device, but then it automatically turns itself off. I usually just proactively make sure it's deselected when I add a new device since it's something you can turn on/off at will anytime. But if left on indefinitely, it can gum up your logs.
For your unplugged stuff, yeah -- it's best to plug those back in. If you have plugs you just use seasonally or whatever and then otherwise stow away, most folks just use wifi plugs for that instead.
I would start with doing firmware updates for as many devices as possible. Some can be done directly if the update is given to HE, and others you will have to track down from the manufacturer and download the package from them.
You did very well to add 56 devices and to only have one ghost. I would suggest removing and re-pairing the one errant device (079) "floater old tulip"
Having moved from Vera to HE over the past 14 months, I have anecdotal experience to suggest the following changes:
Break the refresh rule into multiple rules a few minutes apart.
Avoid scheduling rules on 00/15/30/45 as some apps default their refresh at that those time(s).
Some devices require a 1-5 second delay after a refresh. Instead of one 2 second delay, try two 1 second delays. (I found this to be true for Sengled bulbs for rules to process correctly.)
Device driver selection can be important. Unfortunately the driver suggested by HE may not be the best. (None of my Leviton devices use the suggested driver.)
Rule construction can effect device operation.
In each device, set the event history and state history to 2.
By making too many changes at one time it’ll be difficult to determine if any one change made a difference. Consider a reboot and not making any changes for three days and then see how it operates.
Also, I only use Rule Machine (I uninstalled the Basic Rules) as it seems to process more efficiently. More importantly, however, I learned to write better rules because I concentrated my efforts on learning one tool rather than two.
When I first migrated devices and rewrote rules, I did one every week (+/-). In December 2022 I decided to move the remaining in a few days. That was a mistake as I had simply made too many changes. Everything is working well now, but if I had followed my original schedule (AKA patience), I would have avoided some grief.
What is your zen77 controlling and what is the total wattage? They don’t like really light loads.
How often are you getting the false reports? If you turn off the breaker for 10 seconds and turn it back on, do the false reports clear up for a couple days?
@ FiveNines - I'm going to give all of what you suggested a try.
In the meantime, I'm beginning to narrow one of the problems down. The zooz dimmers are each controlling one 60w incandescent bulb. I've been digging into the logs and events and narrowed it down to one Rule that seems to be causing the problem. The rule Refreshes all the devices every 60 minutes, 5 at a time with 1 second delay between each group of 5. Its a simple Rule only with Refresh and delays. I have it start at 6:15am and at that time the Zooz occasionally respond by reporting they are dimmed to 10% and On even though they are not. Turning them Off from the device page resets back to normal operation even though they are already off. I'm going to put bigger delays to see if that helps but it seems like it should work the way I have it setup.
The problem you’re having is similar to the one I had with Sengled color bulbs. I wrote a rule to cycle red then green sunset to sunrise for Christmas. The color would change (or sometimes not), but the logical device was not updating. Different drivers (I think there are 4+ for those bulbs) and delays seemed to fix the problem. I did notice that setting a level causes problems and requires more of a delay, whereas off/on does not.
Consider putting that device in a rule alone. Rules can be cloned and unneeded commands deleted or commented in the clone.
More delays seem to help (at least after 1 day of testing). Hopefully it will stick. Could be I overwhelmed the network by refreshing everything too quickly.
