Hubitat Migration/Tips Thread

I thought it might be useful for some of us who have moved from Hubitat to HA to have a place to discuss some things that are conceptually different and share tips and tricks for migration. Having gone from Iris->SmartThings->Hubitat->HA, the move to HA was certainly the most overwhelming one for me. I’ve started a list of tips below, and I certainly welcome updates/corrections/additions - I’m still fairly new to HA myself.

I’ll also mention that one handy integration when migrating is the Custom Component: Hubitat which allows you to keep your devices on Hubitat while you learn/evaluate/migrate to HA. No additional code needs to be installed on your Hubitat hub, so if you’re having performance issues, you can run only “official” Hubitat code (though some have reported still having hub slowdowns).

A few other things I found confusing/challenging:

  • There is not exactly an equivalent to Hubitat’s HSM in HA. You have to activate the Manual Alarm Component (which is unintuitively called Manual instead of Alarm) to give you the virtual on/off switch for the alarm system, and then you then have to craft an equivalent safety/alarm system with automations/node-red/etc.
  • There isn’t really an equivalent to mode manager. If your “modes” are tightly coupled to your alarm states, you can naturally just use that. Otherwise, you can configure a Input Select switch and use that state to manage your “modes.” Ultimately, HE basically worked the same way except that it explicitly created that switch for you and and wrapped a little GUI around it. This is probably a bit more flexible.
  • As of June 2020, the Z-wave platform is in flux from an old integration to a new one. Personally, I think the best advice is to keep using the Hubitat hub as a z-wave controller with the above integration until the new Z-wave platform is more complete.
  • The Zigbee component (ZHA) is stable and the majority of devices work, but it doesn’t seem to have quite the breadth of specific drivers that Hubitat/ST have. For example, Zigbee keypads like the Iris/Centralite ones do not seem to be supported yet.
  • For doorbells and chimes there is a feature request that has been accepted but not implemented, so these don’t yet work as you would expect. Personally, I miss this functionality for setting up a door open/close chime. What I’ve done to work around this is create a virtual switch in Hubitat with a Rule Machine rule that triggers the chime, then export that switch to HA. This workaround is likely to work for a number of devices that aren’t as well supported in Hubitat.
  • I still haven’t figured out a great solution for locks. It appears that z-wave locks were reasonably well supported with the old integration, but they haven’t been implemented in the new integration/interface yet. Zigbee locks pair, but there doesn’t seem to be a Zigbee Lock Code Manager, so you can mostly just do lock/unlock/status. Once again, it seems the best option might be to keep those paired to Hubitat until the situation changes.
  • This is largely opinion, but I would recommend not bothering to learn the Automations part of HA right away. You’ll spend about the same amount of time installing/learning Node-Red, and it will probably give you a simpler and more powerful way to manage your automations. Or if you are an experienced Python programmer and loved building apps over on Hubitat, go look at AppDaemon via the HACS add-on.

I also just want to mention that although I’ve never been a very chatty poster on any of the forums, seeing a lot of the familiar and respected names from the hubitat (and ST) forums showing up over here made the migration a lot more comfortable for me. I hope this thread gives back to some others who eventually make the move.

6 Likes

I wrote an appdaemon that you can use with an input_select for modes. I then tie all my automation to the mode, much like HE. Maybe I’ll get it polished up and genericized for the community, when I can.

zwave is a bit of a mess on HA, but when you get the nuances it does work. However, if I could give one recommendation for HE migrants, I’d say new migrants should leave their zwave devices on HE and use Jason’s wonderful HA->HE integration app. Automate on HA, keep zwave on HE just for use of the radio.

I’ll echo, zigbee locks and zigbee thermostats are better left on HE and integrated via Jason’s app.

HA is a bit of a learning curve, but once you get it going on a good piece of hardware, you will not be disappointed.

@omayhemo I still vote you write your presence gov app for HA :stuck_out_tongue: I’ve tweaked that damn bayesian sensor more times than I care to admit, and played with the person integration. And while the bayesian sensor is great at determining presence, it’s still coming in a solid minute slower than your app so if I use it to trigger my garage door I find myself backed up to the door waiting for it to start opening.

For locks, I had moved my Schlage locks from ST to HE and they worked OK, but I found them slow. I then installed a new ring alarm system which supports Schlage locks as well so I moved them onto there and they are faster than I’ve ever seen them! Even better that Ring also integrates into HA and the ring app allows them to disarm the alarm and send a notification on who did it so that took care of both automations I had setup for locks.

For HSM and Location Modes I’m currently left with 2 rules in Rule Machine that turn on a virtual switch in HE and then I have a flow setup in node-red to change an input-select on HA that triggers automations. This is just temporary until Jason gets both of these added to his integration.

My overall goal, which I’m 95% of the way there, is to use HE for the radios and do 100% of everything else in HA.

I’m a few weeks into my “move” from Hubitat Elevation to Home-Assistant

At first I was trying to cut out the Hubitat completely, but I realized the error of my ways as HA’s device driver support isn’t even close to as robust to Hubitat. I was trying to move stuff over from being paired with HE to being paired to my Raspi4 and GoControl Zigbee/Z-Wave stick. While devices would pair to the GoControl, HA simply wasn’t doing a good job on the driver aspect. That’s fine, I have no problem with continuing to use my HE as the device pipeline to my current HA based system. (thanks to the excellent custom component mentioned in the original post of this thread)

So as it stands right now I have 42 IoT devices paired to Hubitat, there are no apps installed other than a single instance of MakerAPI, no custom app code, and two custom device drivers (one for my IKEA Fyrtur blinds and the second for the Sengled 4 button dimmer). The hub responses quickly, hasn’t been rebooted in weeks and does a great job exposing the hardware to HA (and down the line, Node-RED)

Not included in those 42 device are 4 rooms worth of Hue lighting, as I have them talking to HA through the Hue integration to create the entities and then use Node-RED to tie together rooms with Hue + other devices. So for instance the Hue hub turns lights on and off with the dimmer in the room using all it’s native code/app but a Node-RED flow listens for on and off and turns on non-Hue devices with it.

I had moved almost all my rules from Rule Machine to Node-RED prior to getting into HA, so that wasn’t really much a big deal to finish off Rule Machine completely.

HA has opened up all kinds of doors, some literally! Using the HomeKit integration I can now control my Schlage HomeKit lock (using a virtual switch in HA and a HomeKit automation), the Motion Eye add-on allows my RTSP cameras to show up in my dashboard, I’ve got a live running list of aspects of my media server, and things like the WyzeApi integration has works excellent! HA has does a great job of taking integration to the next level but I still see my HE being used for device drivers.

I do miss the “Combined Presence” app on the HE as I’m finding HA’s built in presence unreliable. For things like this, maybe i’ll toy with the idea of installing that app back on the Hub and then using MakerAPI to send the presence to HA.

All in all, I really think my current set up is “it” for me, it’s been really flexible with the wide array of devices I have. “Floorplan” is what got me interested in HA enough to make the move, but I still have not finished getting an image of my house together :frowning:

If I hadn’t already invested in a USB Zwave/Zigbee stick, I probably would have done the same thing myself (though I suspect a couple more HE lockups would have dissuaded me eventually).

I suspect the egos at Hubitat will prevent them from ever doing so, but I really think it could be profitable for them to embrace this model. They keep hammering the gong that they are targeted at simple automation for basic users and that the more advanced stuff is not really their market. If that’s true and not just an excuse, it would be great to embrace having HA users buy their hubs and build off that synergy.

Their business model is predicated on a preposterous concept. There is no such market. DIY automation is a enthusiast’s market.

I have 3 friends/family members for whom Hubitat sounds appealing on paper. I would classify them as novice HA and mid-level technology users at best. They all 3 set up SmartThings on their own, but I wouldn’t recommend Hubitat to them because I don’t think they could get it going on their own and the hub slowdowns make the cloud outages of ST a wash.

I do believe there is a potential market there, especially focused on privacy, but… that’s a very tiny market and Hubitat needs to hire a UI designer and abandon Rule Machine immediately if novice users are the target (“we’re on the 4th major version of something that’s not intended for our target customers”).

I think what it comes down to is that hub-based automation is currently only necessary for more advanced users. Hubitat set out to steal SmartThings’ user base, but meanwhile the tehcnology has shifted. Most novice HA needs are served by installing Caseta/Hue/Ikea, Ecobee/Nest, and Google Home/Alexa - Hubitat would be redundant.

Exactly. That market (as tiny as it is) is Hue and such. No additional hub required.

I couldn’t agree more. I consider myself an advanced user. I know a lot of people that are getting into home automation that are nowhere near advanced. And every one of those people have the cheapest devices they can find and are extremely content with telling google/alexa to control it. That’s all they want or need. So to build a business with those customers as your target market, you’re going up against the giants and will get squashed.

It’s good that they allow devs to create apps, but to point the finger at those apps when things go to shit and blame them for all the issues when it’s quite clear that the hub itself is under powered.

Hey guys, I think I may migrate is I ever figure out the hubitat integration. I’m already in my head “give me a reason to buy a synology nas with docker, I’ll do it!” lol

Anyway just came to say hi

The HE to HA integration is super simple. Simply setup MakerAPI on HE and choose all the devices you wish to have in HA.

On HA
Install HACS
Install HE Integration using HACS

It’s that simple. The devices in your MarkerAPI will show up in HA.

The problem is installing hacs, I’m on a vm and I donno what’s going on lol

So, no supervisor?
Didn’t want to assume you’ve seen the documentation.

I have supervisor, I did hass.io

Is there something in the HACS docs you’re getting hung up on?

My biggest hang up was deciding which version of HA to install.

Side note: decommissioned my last HE

:confetti_ball: I’m free of Hubitat :confetti_ball:

4 Likes

Hubitat announced a new hub today with the only improvement being a new z-wave radio and z-wave software stack. My Hubitat slowdowns always followed my z-wave devices across my 3 hubs, so it makes me wonder if they tracked down the problems and this is the fix. They didn’t bother upgrading anything else, including the zigbee chip which is a bit outdated, and it’s not like the 500-series chips are dinosaurs or unavailable - they spent the money re-engineering just for this one component. It just really jumps out at me as being… “interesting.”

So many word choices about that announcement have me thinking that either things aren’t going well in Hubitat land or that they’re about to finally unveil whatever they have in mind for profitability.

Umm what re-engineering? They just upgraded the z-wave board which now includes the 700 series vs 500 series chip… there’s no engineering involved it’s just a chip swap. What’s to be seen is whether or not they have actually updated the z-wave library in their software stack AND fixed some of the problems in their software stack.