Slow web interface

🛎️ Get Help
143

63 is a 'moderate' sized collection to my eyes. Naturally it depends on the types of devices. 65 Power Monitor outlets is massively different than 65 door sensors. Look at live logging and watch for 'bursts' of logs. I personally have a upper limit of 4 'cause-and-effect' events per second. That is, one motion event causes one light on/off event = 1 cause-and-effect. Currently I run at about 2 seconds per 'cause-and-effect' events... so quite far from my imaginary limit.

Are you seeing page display slowdowns or 'cause-and-effect' slowdowns or both? There are different workarounds for each.

144

I have about 5 power monitors (Zooz Power Switch and Zooz Double Plug) and I have noticed them very chatty in the log even if in the device it it mentioned to report every 30 seconds I see in the log the power many time per seconds. Let me try to remove the logging all together (they do not have debug logging, just text information).

145

Maybe a little bit better with less logs from the power but still slow first page loading. For example, going from devices to apps:

image
image1257×458 143 KB

146

Try disabling the power monitoring features of these devices at least temporarily to see if it helps to calm down your Z+Wave mesh network. Many users have had excessive power monitoring data cause hub performance issues.

2 Likes
147

Thank you for your suggestion. I will give it a try.

However, I am surprise the the Z-wave network has an impact on the web interface. The web interface should get its data from the database while the Z-wave performs its action on a queue/bus. They shouldn't impact each other. That being said, I have notice some "lag" in few lights/sound of 2-4 seconds but nothing as slow as the web interface. I'm not sure what is the architecture of the web application of Hubitat but it would benefice a lot of a good separation (if it is not the case).

1 Like
148

Slow UI (web interface) has other factors that have been raised. There's a speed/duplex (100/full) that is supposed to have a patch applied to the Hub automatically in 24 hrs of detecting. There's also another conversation about CRC errors on the wire. If you have a managed switch you can determine if this applies to you.

But when push comes to shove and there's a lot of events on the Mesh, I am not surprised that the UI takes a back seat.

149

I don't disagree... However, every one of these Z-Wave power updates will generate a unique Event for its particular device. Every one of those events gets written to the database, and every App that has a subscription for those events, will be instantiated and executed. So, I personally believe high frequency data from devices can impact the hub's overall performance.

1 Like
150

I do not have any insight into the Hub itself, so I can only talk from previous system experience I built or had to work with. Because Hubitat is all local and also has one of its selling points to be fast, it is a little disappointing (as a customer) to have 10-25 seconds page navigation.

Eventing prioritization is hard, I am not diminishing the humongous amount of work already in place. But, it makes me think that maybe two separate servers (one for Z-waves communication and one for the interface) may be needed if the Hub itself is overcharged.

I see the hub as a web application server that receives hundreds of hits by a user. The write and read need to be separated to improve the user experience. The writes cannot block the read, caching for reading operation on the webserver, etc. Even on high read/write, there are strategies to mitigate slow user navigation on the UI.

But, I'm new to this system, and I do not believe anything is open source so, I probably won't be able to do much :slight_smile: That being said, I like few things but the rule "language" combined with the slowness of the system starts to irritate me. Maybe it will go away once I have fewer rules to configure. Concerning the rules, if I could "code" somewhere else and just copy-paste, I would be less impacted by the slow web interface, but, so far, it seems that I'll have to live with it.

I'll start trying the suggestions! Thanks

2 Likes
151

If you're programmer, you can write your own apps in Groovy. This provides the ultimate in flexibility, but also allows one to screw things up as well. :wink:

152

10 to 25 seconds is not normal, at least not by my experience. I have sub-3 second times either on cloud or local.

I have over 130 devices (including a couple power monitoring), a dozen or so Simple Lighting, a dozen or so Motion Lighting, 20-ish Rule Machine rules (some with dozens of lines of IF-THEN statements), plus maybe a dozen or so more assorted apps like Better Laundry Monitor, Echo Speaks, Rheem, Honeywell, and so on.

In my opinion, there is something about your setup, either an app or device, that has a problem.

1 Like
153

I'm one of those people who had issues with too many power updates. It could have partially been a result of my older 300 series Aeon v1 HEMs, but the most likely reason is that I'm the one that caused the issue by setting one of them to update every second and the other to update every 5 seconds.

Since moving them to another hub, the slowdowns are gone from my main hub, and the slow down after two days has now moved to my second hub, which has the 300 series Aeon v1 HEMs on it, and only those two devices. I'm not a heavy Z-Wave user like you are, but the two Z-Wave Plus devices that remain on my main hub are as happy as ever, as is every other device and automation on my main hub, as am I :grin: Gone are my slowdowns that previously started after just two days.

A 15-20 second delay would not be acceptable to me either, but I can assure you that this isn't just the way things are. There are many other users with as many Z-Wave Plus devices as you have and more. They do not experience these slow downs, so you have something currently dragging your hub down. I can tell you that it is quite a nice feeling when you finally find it, and the hub becomes very fast and quite enjoyable to use, so I do hope you discover the cause soon.

Your concept of the WebUI and rules on one hub with Z-Wave on another is in fact what many have done. This isn't necessarily a requirement, but it can alleviate the problems that devices can introduce. It's quite difficult to add many different devices from different manufacturers, and not have anything go wrong. They can all follow the Z-Wave spec and all have certification, yet any single issue in their firmware can cause a chain reaction that can mean devices just need to be rebooted, reset, and sometimes retired (or exiled in my case) if they're causing too many issues.

In regard to rule machine rules, unless you're using global variables (which is known and was stated to cause a performance hit), your rules should all be fast and there should be no degradation regardless of the number of rules. I have over 100 rules, and I have no slowdowns from it.

I cannot discuss what is in the beta testing group, but do know that the Hubitat team is aware and are continuing to address these concerns.

2 Likes
154

fwiw, my slow interface issues went away completely when I replaced my Schlage locks (Bobby said one of them was flooding my HE with bad data), and the 2.1.9 update. Been completely reliable since. Had to reboot the hub to get zwave pairing to work the other day, but that's happened before where pairing just doesn't work unless I reboot the hub.

4 Likes
155

I've turned off all my Zooz Double Switch and Zooz Power Switch off in term of log

image
image1049×113 2.53 KB

I have a Schelage, but I do not see any log. I might try to exclude this weekend... but at the price it cost I really hope to keep it connected. That being said, I have an automatic rule at 10pm to lock and the last two days it didn't lock.

I've notice the Thermostat missing some time scheduled temperature, noticed lag on rules (e.g. chiming on door open, or opening light on motion).

Is there another way to activate some more advanced diagnostic tools on Hubitat? At the moment, it looks a lot like trial and error...

156

I don’t think just disabling the logging will have a significant impact. If possible, I would disable the power monitoring capability of the devices via their device details pages (if possible.) This is the only way to calm down the Z-wave mesh to see if it has an impact on your hub’s performance.

4 Likes
157

Let me check that out. Not sure it is possible.

This is what I have at the moment on the Zooz Power Switch:

image
image1628×639 24.9 KB

This morning, the Thermostat didn't adjust. It seems that everything is getting worse and worse. Now I am reaching over 30 seconds on the web UI.

image
image1236×595 35.9 KB

I will pause most of my Rules. I suppose one of it might impact? Not sure...

158

It looks like it is possible. Try setting all of the energy/voltage/current reporting to “No Reports”. Do this for ALL of these devices to try and calm the traffic down.

image
image1668×1871 362 KB

1 Like
159

I will, but wouldn't that make the switch completely not smart? Also, I see in the log many report every second. I am confused about the 30 seconds interval from the UI when in the log I can see several entries each minute.

160

Also, I noticed that every time I reboot the Hub, it is fast for a couple of hours (about first 2-3 hours). Then, with time, the performance degrade.

1 Like
161

The switch will still turn on and off via Z-wave. We’re just trying to see the impact of excessive power reporting on your Z-wave mesh network/Hubitat Hub.

Once we know what the impact is, we can start troubleshooting Why? And then the How to fix it. It may require assistance from Hubitat Support, if the problem exists in their Zooz driver.

Also, when changing the settings for the devices, be sure to click SAVE and the click CONFIGURE. This will ensure the physical devices receive the updated settings.

2 Likes
162

If you have not already, you might also try a factory reset of that outlet, then general exclusion, wait 30 seconds and rejoin it.

1 Like