Losing patience with hubitat slowness

OK - I only published the pre beta2 link on Tuesday. It is significantly faster than beta1.

Couple of things... create a standard virtual switch device in HE (not in my app) and publish that to MQTT and discover in HA. Then we can eliminate any Z network delays. Time against that.

The onoff control switches in HA that use optimistic/predictive state should now not move say ON then OFF then ON again. A behaviour you saw before.

You're not using HA Discovery (statestream) into HE I believe but that's not a concern anyway.

If you need any help or have observations I'm around all weekend.

K

Absolutely right but I'm fairly confident though... I have a very event intensive MQTT setup and I do not have issues, and I have a user with nearly 500 nodes in MQTT that is not reporting issues. Fingers crossed. :crossed_fingers:

Oh...I make no claims either way. I'm just saying, and you have to agree, that the echo skill's behavior isn't evidence either way.

I think that's what I said..

oh... maybe I am on beta 1 then. I believe I'm using:

Alright well I have a nightly reboot strategy that's working for me without leaving my garage door open.

from the linux box that's running my appdaemon container:

00 04 * * 1,2,3,4,5 /usr/bin/docker stop appdaemon
01 04 * * 1,2,3,4,5 /usr/local/bin/restart-he.sh
20 04 * * 1,2,3,4,5 /usr/bin/docker start appdaemon

where /usr/local/bin/restart-he.sh consists of:

#!/bin/bash
curl -k -c /home/bill/hubitat-cookie  -d username=reboot -d password=reboot https://192.168.xxx.xxx/login
sleep 5
curl -k -sb /home/bill/hubitat-cookie -X POST https://192.168.xxx.xxx/hub/reboot

The 20 minutes before restarting appdaemon give hubitat time to reboot and MQTT publish time to complete. It seems to take about 10-13 minutes in my limited testing.

In another thread, there is some suspicion that one of the causes of slowdown seems to be CRC errors on the NIC of the Hubitat.
Is there some way you can check if that is your issue as well?
In that other thread, various individuals were having transmit/receive errors causing a slowdown in the data via ethernet. They were able to see that because their Hubitat was on a "smart" switch from which they had some statistics.
If nothing else, see if you can check that out.

2 Likes

So what is the solution to the CRC errors? My hubitat is not on a smart switch so I can't really check.

There is no solution because it's still being investigated (maybe). But the chance you have a faulty unit is very small. The network crc error would make the UI slow but does not effect the zigbee or z-wave that much to notice. So I think there is something else going on.

1 Like

Same here, hubitat becomes very slow and requires reboot. I think we need to post more in this thread so developers will pay attention and provide ways to debug it.

Web GUI access stuck on
image

Which for me would be indication that app run out of sockets, which points to WiFi based devices, and I recently added two:
Sonoff switch
Maker API

Could also be google home's integration fault...