Severe Hub Load with Hue Bridge Integration

Hey everyone,

Having an issue here with the Hue Bridge Integration, the built-in one from Hubitat, specifically. I had previously used CoCoHue without problems, but seeing as it seems to have become integrated I thought I may as well move over last time I was doing some messing around.

The following week I noticed issues with the Zigbee radio dropping off on my C7, and found it reporting severe hub load, with the Hue integration being a high percentage of usage. So I removed the integration from the C7 and installed it on my C8, which has fewer devices attached. The problem moved, and now my C8 is experiencing severe hub load issues.

Trick being, there are only 21 lights on the Hue bridge. No other accessories, no cameras, no security stuff, so this shouldn't be the hard work for the hub that it appears to be, surely?

I offloaded all of my lighting (except for in-wall on/off modules) to Hue years ago as I found lights that provided Zigbee routing being turned off would really destabilise the mesh on Hubitat, and this has been a solid compromise ever since.

Can anyone offer any troubleshooting steps or insight into why this should be happening?

Sorry to prod, @gopher.ny, but this is happening almost every day now. Any chance someone could give me a hand with this?

I wonder if it's your Hue Bridge that's causing the issue. I and many others have had the Hue Bridge Integration running for many years (Since my start with Hubitat on a C3) and I've never heard of the Hue Bridge Integration causing severe hub load issues.

Are there third-party apps, or Hue Labs apps running on your bridge that might be responsible?

Not a sausage, it's as plain and boring as a Hue bridge can be.

That's the funny thing, I've run this for years without issue with CoCoHue. Since switching to the integrated version it's been hitting >50% busy time on device stats and >80% on app stats until eventually the hub stops processing automations. Leave it any longer and it shuts off the Zigbee radio. :man_shrugging:

I find the "% of busy" measure of load weird too, because surely a hub can be virtually unused and something will take up a significant percentage of that very little work. I only know the hub is busy from the warning and if I check /hub/advanced/freeOSMemoryHistory. Right now it's:

Date/time,Free OS,5m CPU avg
02-18 22:38:25,637952,1.76
02-18 22:43:25,475072,2.28
02-18 22:48:26,450536,0.44
02-18 22:53:26,425632,0.32
02-18 22:58:26,391664,0.3
02-18 23:03:26,351956,0.48
02-18 23:08:26,329332,0.38
02-18 23:13:26,321660,0.4
02-18 23:18:26,319344,0.38
02-18 23:23:26,318684,0.44
02-18 23:28:26,317012,0.29
02-18 23:33:26,317160,0.4
02-18 23:38:29,348408,0.48
02-18 23:43:29,342920,0.28
02-18 23:48:29,342100,0.44
02-18 23:53:29,339368,0.43
02-18 23:58:29,337720,0.4
02-19 00:03:29,331572,0.53

But towards the end of the day it's up to >2.7 on the 5 minute average. The hub reboots every night already, so this is accruing only over a maximum of 24 hours. :frowning:

I can't imagine it would be related to hub load issues, but are the Zigbee channels at least semi-separated?

My Hue bridge is on CH 20
Main Hubitat is on CH 13
Secondary Hubitat is on CH 25
Home Assistant with a ConBee 2 (Z2M) is on CH 15
Home Assistant with an HUSBZB1 (ZHA) is also on CH 25 <- I know this is wrong and I wasn't paying attention when I installed it. Ironically it hasn't been causing an issue. There's just 6 Lutron Connected Bulb Remotes on it. I'll change it when I notice a problem. :grin:

My Hub bridge just has 29 Hue bulbs and 3 Hue Outdoor motion sensors joined to it. Is yours filled to the brim? I have heard of performance problems when you start to approach 50 devices.

1 Like

Andy, just asking about your sausage:

I don’t see that you have mentioned your Hubitat firmware version. I saw the same thing beginning with 2.4.0.145 and for a few releases following, now cleared up (2.4.0.151). You wouldn’t by chance have one of the problem releases, would you?

Yup, they're all fairly spread out. I feel for my neighbours. :joy:

Nope, just 21 lights. Nothing else.

Nope, all up to date on 2.4.0.151 here. :confused:

Now I'm thinking about sausage sandwiches, bit of ketchup, crispy onions... :slightly_smiling_face:

2 Likes

Okay, I've got what is possibly the worst "solution" to this, in that it just kicks the can down the road, keeps me digging the hole a little deeper, mixes my metaphors, and doesn't actually fix anything.

I scored a deal on a C8 Pro, so I'll migrate from the C8 when that appears and see how long it takes to fall over again, then wind up back on this thread.

These severe load issues are really the worst problem I've had with the Hubitat platform over the years. I've found it crops up far too often, and I'm really not doing anything intensive on these things. I even split my gear up between a C7 and C8 to share the load, which is particularly important given the AlertMe / Iris devices I have installed, which aren't supported on the C8's Zigbee radio.

The worst bit is when the load gets too high the hubs shut their radios off, which may not have been the source of the problem. When this happens I always have at least one device which panics because the coordinator is gone, leading me to go digging behind furniture to reset and re-pair dropped devices.

What a stupid hobby this is. Nobody hates technology more than people who love technology. :joy:

Wish me luck, see you back here in a week! :wink:

1 Like

I have same issues with my C8 and Hue. Any other thoughts for a solution? I have a retired C5 that I could bring back into action but not sure if that'll help...

1 Like

Given the nightly reboots, my initial sense is DB corruption - Can you do a DB rebuild and log purge, followed by 20 min power off (to trigger a Zigbee client panic, and let Zigbee routes get resorted)

Then temporaily disable the nightly reboots..

As for Zigbee radio dropping, severe load would likely do that, as Zigbee is usually the 1st radio to go down - But have you looked at "real-time" Zigbee traffic, on Settings->Zigbee Details->View Logs just to make sure there isn't some Zigbee device spamming the mesh, which is also contributing to your load issue..

Just some initial thoughts - YMMV

I'm of no use to you, I'm afraid. I'm glad, but sort of frustrated, that since moving to the C8 Pro the problem has vanished. I've re-jigged everything so that the C8 Pro does practically all of the work, with my C7 now essentially being a relay via Hub Mesh and HomeKit for all my AlertMe devices.

The really confusing thing is that the C7 did the job perfectly for years with no issue. It may just be coincidence, but it was after moving from CoCoHue to the native integration that I began having problems, but even then it didn't begin immediately. It took months to become an issue, so that the native integration was possibly responsible is pure conjecture.

I did try the DB rebuild, though not the 20 minute power-off. Not brave enough to do that. You just know that the least accessible devices would refuse to rediscover the mesh afterwards. :joy:

Yeah, the AlertMe smart plugs are quite chatty, but they have been for years (and they're really not that bad) so I think I can rule out Zigbee traffic. Thanks for the suggestions though!

Apologies for opening up an old thread, but I started experiencing this exact issue about a week ago (first with CoCoHue, then I uninstalled and added the native Hue integration. Same issues).

What happens is my C-8 will just completely lock up. After power cycling I see the warning about severe load, and the Hue Bridge integration is basically the entire load under both drivers and apps.

I've tried what was suggested earlier in this thread (rebuilding the database, powering down the hub for 20+ minutes) but it seemed to have popped up even faster today. I've since disabled the integration, but I'm curious how this only suddenly started showing up and I've had the C-8 (and the C-7 before it) with Hue since launch.