I've been looking at the larger number of integrations that exist for Home Assistant and how much more flexible their UI Dashboards can be at the same time as considering trying to get the best of platforms outside of HE (e.g. Node-Red) and the pros and cons of this approach.
I'm thinking that if I used HE purely for Zigbee/Z-Wave device managment, Home Assistant for Dashboards and then something else for logic/rules/automation, which could be any of these 3 platforms, I'd get flexibility, diversity and maybe the best of all worlds?
Anyone care to chip in that has gone this route with their experiences?
Pretty sure a number of user have gone this route (there are a few threads in the HA community).
At this point, I have pretty much everything going through MakerAPI and node-red and I'm a happy camper. Moving everything off of the HE as well as doing away with as much Z-Wave as possible has made my hubs rock solid with none of the forced reboots that plagued it before. I'll be honest though, I'm not sure what changes made what difference when you throw firmware updates in there as well.
I have very little use for dashboards with device controls so I didn't add HA into the mix. Between HE and Node-Red integrations, I've run out of things to automate...
Edit: I'm not knocking HE's implementation of Z-Wave. I just didn't have any faith in the Z-Wave switches that I had installed and preferred using Lutron so all of my Z-Wave switches got tossed into the lessons learned pile. I still have some outlets and Zooz ZEN relays around.
I have been using Hubitat with Node-Red for over a year now. I see no way of going back to using any other automation system.
I very recently added another Pi running Home Assistant to take advantage of the Local Tuya integration and to be able to control my GE wifi air conditioner. I integrated HA into my Node-Red and everything works perfectly for a couple of weeks now.
I use Node-RED on an RPi4 for all my rules and to tie in my 3 hubs, a personal weather station (mqtt), and HA as a test.
Node-RED on a "companion server" is the way to go. I use it for my home setup and my residential clients. I also use the Node-RED Dashboard as well. You can also install other things along with Node-RED like Homebridge, Wireguard etc.
I also installed HomeAssistant on an old RPi3 and am testing a few things for fun. Due to the finicky nature of HA I would probably never install at a client's - it's too support intensive for my tastes. Having said that everything I did worked fine - paired some devices using my old C-4 Nortek stick, tied things back in with HE etc. It does have a nice dashboard and the available plugins are very extensive.
I hear of people using Home Assistant in much the same way as Node-Red.. as the automations engine. It also gets used for some wifi/LAN uses, but the stories regarding ZWave and Zigbee still seem challenging.
I have had that experience. My most recent excursion into Home Assistant was 5 months ago now and I chose to use that Nortek USB for ZWave and Zigbee. I never tried Zigbee at all but I had limited success with ZWave. They had just brought ZWaveJS out and I rebuilt my working system to use that. I did eventually get the 3 ZWave devices working again and it was really reliable. However, each device type I added, and of course the 3 I added were each different, were a lot of fiddling to get working. Once over that line though, it was solid. I included a recessed door sensor, an appliance module (wall wart) and an RGBW bulb. Opening and closing the Door turned on and off the bulb and appliance. I simply could not see myself working through the dozens of different device manufacturers and vintages to get a reasonable number of devices directly connected.
I got it working into Node-Red and that really made me treasure the Hubitat interface. It's been powered off for a couple months now since there's nothing I have that needs it's integration options.
Well to be fair HomeAssistant has a version of Node-Red you can run as a plugin... but I completely agree with you.
If someone is really into the whole DIY thing and wants maximal control over everything and doesn't mind hacking obscure config files from time to time or chasing down regressions after an update then HA is definitely for them.
I did set up HA for a friend of mine on another RPi3 with no devices and just the airplay and ?chromecast? plugins. Was kind of cool because he can stream to his old chromecast audios plugged in to various receivers using his iPhone.
I've decided to increase my use of Node Red for rules, since HE continue to develop their rules engine but requires us to recreate all rules where you want to make use of new RM version functionality. That for me is a real issue and an unsustainable approach. If other software companies expected you to rewrite your documents, slideshows, music, engineering designs etc at each major release of software with no way to simply open those existing data and extend them with the new functionality, well they wouldn't be around for long. Take for example the new hub variables that correct the mess of what was created previously in RM. Now to migrate to the new hub variables is a major rewrite of many of my rules. I started and quickly decided I just don't have the patience to do this anymore. No thanks. I'll move those rules to Node Red instead and gain that flexibility, security, independence, efficiency and definitely the performance improvement.
For dashboards, I do use HE's dashboards but they are so limited. No drag and drop, so a very clunky set-up. They also look so poor, I just don't like grids of tiles. I do agree that HA has quite nice dashboards, especially in style. But having to cope with HA, all those breaking changes, another layer of complexity outside of my control, it's too much. I'd rather just build my own dashboards using html/css/javascript and have complete freedom to create exactly what I want. It actually works great with HE using Maker API and the eventsocket. Horses for courses. Many on here are moving to using HE as a zigbee/z-wave device controller.
To be fair, things like Rule Machine are completely different apps between say version 4.0 and 5.0 (and to some extent firmware are this way too). It would be like expecting Wordperfect for Win3.1 to run on Windows 10.
You don't have to move anything if you don't want to, and you really shouldn't have to move that many apps to a new version.
I'm yet to see a ~$100 platform that works the way you would like. I use zigbee2mqtt extensively. There have been breaking changes twice in the last year, so I haven't installed the last two updates. I used ZwaveJS for 7-8 months a while ago. Every other release had a breaking change, compelling me to stop using it. I used Home Assistant for a year, until a breaking change killed all my existing automations.
Old Vera users will recall the "sh*t-show" that accompanied moving from UI5 to UI7.
Compared to all that, Hubitat platform updates permit EXISTING automations to run exactly the way they have. In many instances, years after Rule Machine updates. I believe @csteele still runs some original RM automations and some RM 2.5 automations.
I understand that you want a migration path where you can import a Rule 2.5 automation into RM 5.0, and take advantage of RM 5.0 features. I'll bet that making that happen will balloon the RM 5.0 codebase, with substantial consequences to its efficiency - although I should let @bravenel comment on that.
I don't expect that from any appliance that costs ~$100. Heck, I recall spending $2300 for a Gateway 2000 386 in 1990 that ran Windows 3.0, and Word for Windows 1.0. That utility of that computer lasted all of about 3-4 years. It wouldn't run Windows 95, and whatever version of Word accompanied it.
I am raising this because several people have raised the analogy of newer versions of Word being able to tackle documents created on obsoleted versions of Microsoft Word. That's true, and ignores the fact that these newer versions of Word (or Windows for that matter) run on bloated, expensive, hardware to accommodate (or permit) backward compatibility, and even then has a usable life (in terms of supporting newer software) of just a few years.
Hubitat hasn't done that. The latest platform continues to run on their oldest hardware. And rules created with the oldest version of Rule Machine continue to run on the latest platform.
/rant off
Edit: I have to edit this because one of my collaborators read this over my shoulder and started laughing. He reminded me that we cannot exchange files between the same version of Word on the Mac vs Windows, without fonts getting mixed up (especially Symbol), arrows getting moved etc. So hey, maybe Word is a bad analogy after all .....
Nonsense, again! Nobody requires you to change existing rules that work. You just want something completely unrealistic, and can't resist beating this old horse of an argument.
LOL. Good luck with this. Time for you to move to a better platform, one that lives up to your expectations.
I’ve been using HA for awhile now and would highly recommend it as a campion to HE. It adds a lot of functionality to my home automation. Even if you just use it for the built-in add-ons it makes it worthwhile. I’m using it for Node-RED, WireGuard, InfluxDB, Grafana, MQTT, AdGuard and AirCast to name a few. Having all these integrations in a single application is nice for maintaining updates. I’m running HA on an iMac and wouldn’t recommend trying to use all these add-ons on a RPi with an SD card, but it should work if you’re using a RPi with an SSD.
HA has some awesome dashboard capabilities that I’ve toyed around with but my wife prefers Apple’s Home app on her iPhone so I’m using HomeBridge to integrate into HomeKit as well.
I think this combination is very powerful, but I have to warn that with that power comes the ability to put almost an endless amount time to do whatever you can dream up. Here is a post I made over in the node red forum:
TLDR, I implemented a flow that takes inputs and shows status of a smart outlet in:
Node Red Dashboard
Hubitat UI from actual device hub
Hubitat UI from virtual device hub
Alexa Voice
Alexa App (pc + phone)
Google Home Voice
Google Home App (phone)
Google Nest Hub Max
With the ability and a clear path to also add:
Node Red connected UIBuilder based Dashboard
Home Assistant UI
Apple Homekit via Node Red to Home Assistant to Homekit
Mostly just a proof of concept. My long term plan is create a virtual device for each place I want control. Then build the interface for dashboards to whichever interface (s) the family likes. Finally this will allow me to things like place one brand of bulb there, then be able to replace it with a different brand or interface say zwave on Hubitat to zigbee on z2m and only have to replace the interface for the specific bulb while leaving logic and interface name unchanged.
Not at all. Maybe you didn't read what I wrote or maybe didn't understand it. I didn't say that I'm required to change any existing rules that work. I stated it is necessary to recreate existing rules in order to take advantage of new functionality, which is absolutely correct. And not an issue for many people with simple rules. And not a problem for many of my rules which are simple. And no longer an issue for me at all because I'm moving all my complex rules to Node-RED.
Not sure why you are so salty about perfectly accurate feedback. My "completely unrealistic" expectations are now completely satisfied with a rule engine that continues to allow me to access my rules through successive revisions of their software, taking advantage of their new functionality without having to recreate anything, unlike yours.
LOL. Good luck with your platform. I'm not moving to a better platform. I'm happy to put HE into its most capable role, as part of a system where it's clear other solutions are better at some things. Sorry if that offends you chap.
If you're talking about node-red, then be aware that there were breaking changes there also. For example, it is planned for Node-RED 2.x to drop support for node 8.0 and 10.0. There are/were also breaking changes between node-red 2.0 and 1.3.x.
And their published release plan specifically mentions the possibility of annual breaking changes:
Yeah but even with those breaking changes they will not necessarily require you to replace your sequences entirely to get new functionality which I think @Angus_M was getting at.. but not trying to speak for anyone!
While not the perfect setup the fact that HE allows older rules to coexist with newer ones is still a pretty decent/cool thing to do imho and as a guess probably sufficient for most of HE's target market.
It's been powered off for a couple months now since there's nothing I have that needs it's integration options.
