I used to run HA on an RPi, but eventually migrated it to a similar NUC. The RPi eventually just wasn't powerful enough (peak compute needs), while the NUCs are still quite cheap. And you can run a surprising amount of Proxmox VMs and LXCs on barely a few cores and gigabytes of RAM.
Cost-wise, there's a solid chance that the Pi would have been more expensive. Jeff Geerling ran some numbers (^1) on this last year, before the current chip crisis we're in, and it was bad enough already.
Home Assistant does a surprising amount of Disk I/O, if for nothing than for logs. Sibling commenters are also advising not running it on the SD card to avoid wearing it out, so there's definitely some truth here. This means we're adding a Pi M.2 hat + SSD into the mix. The Pi5 SSD kit for 256 GB, when it was available, was around $60 USD. A Pi5 with 8 GB of RAM is $130 USD. Now we need a cooler, a case that will fit the Pi5 with said M.2 hat, and a power supply. We're already well north of $250 USD, encroaching on $300, and we're not even using the core benefits of the Pi's platform. No need for GPIO pins, tightly integrated cameras or other sensors, none of that.
For all we know, the blog author did this assessment (or trusted the assessment of others, eg: Jeff) and came to the came conclusion - it wasn't worth the price of entry.
I wanted home assistant compatible plant watering solution that works on a solar panel and does not require being connected to the water line and is Zigbee compatible. Unfortunately, I could not find any. So I did a DIY solution: a big barrel which I manually fill with water, a 12V pump (usually sold for camper vans), some rechargeable batteries, 10W solar panel, a solar charging controller, and Tuya ZG-2002-RF switch.
There are also irrigation controllers that use a ball valve and will work fine under gravity pressure e.g. raised rain barrel. Powered via battery that lasts months and months. No home assistant of course but water usage can be estimated from expected flow and programming.
Fun project. Though it’s kind of unreal how complicated it is to set up HA and I literally do this for a living, both embedded sw and backend web dev.
Docker compose with a zwave management server, reverse proxies for TLS, vlan isolation for the server, macvlan for HA container so it does see the host network, etc, etc. All to turn on and off a lightbulb with the sun. All the while AI is telling me to configure things insecurely.
I think when I get some more spare time, I’d like to write a statically linked program that handles a zwave controller and basic automation scripting. No IP networking needed for my lightbulbs. Then it wouldn’t feel risky to just make a system user and udev rule to give it permissions to the controller, and run with systemd.
I went through a similar process with Home Assistant. And the kicker is that months or years down the line, you'll hit some feature that doesn't work with the Docker version (I've ran into a couple)
I've tried using HA a couple years ago and gave up. It was too complicated to run it in a Pi4 - I'm an experienced software engineer, familiar with containers and Linux.
I was trying to get some of the IoT I have at home like pool equipment, lights, HVAC, blinds, etc. Some of the setup were an uphill battle looking for more information in forums and trying to figure out what was broken.
Recently I decided over the weekend to use Claude and write a small app that controls my pool equipment and then deployed it using Cloudflare Zero Trust (kind of a reverse VPN). What a joy! Not only I had lots of fun reverse engineering my pool equipment API (I didn't want to depend on existing libraries - which I know exist) but I managed to create a fun and custom UI with React that my kids and wife love using. For example, whenever the pool heater is on, it adds an animated flame to the UI and change the background to a red-ish color. Plus it has a bar chart that shows the pool temp progression (takes hours to heat it up) with an animated volcano colors. The theme of the app is beach/pool vibe.
I don't think anyone here would be that excited if we were using the lower-denominator that HA turns out to be. I know it's a very cool automation tool, but just not very exciting and pretty obscure to configure every equipment I have at home.
I've been thinking about writing a blog post with the details of my fun project, let me know if anyone is interested in this. So far I've done the blinds and pool equipment. Next will be HVAC and lights. Took me 1-2 weeks total for each using Claude in my spare time.
> Though it’s kind of unreal how complicated it is to set up HA and I literally do this for a living, both embedded sw and backend web dev.
I had the same thought after I joined a local group for Home Assistant users.
Everyone always talks about a happy path where you pick the right choices, use the right setup, and everything just works immediately. More often when people come to this local group's shared Slack channel it's because they're 10s of hours into trying to set up something that appeared to be simple. Then all of the old timers remember that they, too, suffered through something similar once and share what they can remember.
I think HA can be a lot of fun for people who like to experiment and debug, but if you're not the kind to be entertained by debugging your home's operation then it can feel like a chore. Some have an easy time setting it up and then get trapped when an upgrade breaks something or they try to add a new device with less than mature support.
HA on R-pi running for almost a decade without issues here, including moving house a few times. Sounds like you're making it difficult with that setup. Mine is connected to light switches, alarm, duckdns for outside network access, motorized blinds, garage door opener, hvac, landscape lighting. It's magically awesome and takes none of my time to maintain.
I spun up a VM on my proxmox server and loaded HAOS. Job done. The only added steps involved giving the HAOS IP to my reverse proxy, now it's reachable from the outside net.
I have been using HA to water my garden for 4 summers. I setup a Tempest weather station this fall, and will have some fun experimenting with using rain and temperature data collected in my back yard to make watering decisions.
I love HomeAssistant, and my second time on a new home, i'm slowly getting what i want in terms of interface and devices; doing it slowly helps you plan better and execute it perfectly.
I've also been watering my garden (sprinklers) and i even built my custom ESPHome device: https://github.com/mgarces/open-esp-sprinklers
After years of doing this, I determined that dumb controllers are superior for my uses. A once a week irrigation with a simple rain sensor results in the same quality lawn/beds and better can run with pretty much zero maintenance burden on the sw/hw side. The vegetable garden is even dumber, just drip lines, a hose bib, and a dumb timer that flips daily in the morning powered by AS batteries. In my apartment days I just used an elevated water container and the same dumb timer to gravity feed my garden.
I find home lab stuff has far more return on investment for like automatic blinds, lights, etc. It’s not like you can just stay inside anyways and get amazing vegetables, you still need to be on top of thing like checking for pests and disease. The automatic garden is a myth.
I love HA, after setting it up I used an old tablet as the dashboard to control and monitor the house, plus the app. It’s very easy to setup and work with, maybe a bit convoluted in the initial setup but once done, you barely do anything until you add new sensor. You can setup a whole SCADA-like system with it, controlling your garden, power grid/solar and monitoring it, integrating CCTV and all, and it’s free. A similar industrial project I did before, SCADA and RTUs that monitored and controlled many actuators and solenoid valves and sensors, cost wise was ~$9M, and all the functionalities were implemented can be done with HA on principle.
I use Irrigation-V5 in combination with a flow meter to manage my sprinklers. Irrigation-V5 basically tracks the amount of moisture in the soil based on humidity, the amount of sunshine, temperature, rainfall, and of course when the sprinklers turn on and add water. It feels like magic. They can go for months in the winter without running, all on their own and then pop back on when things warm up and dry out, slowly decreasing the intervals between watering as summer approaches.
Really makes me want to integrate the whole thing into an ESP32 with a display so such a thing doesn't require HA.
I made such a thing ten years ago with an ESP8226 and a basic iOS app built using HTML and javascript. It still works perfectly.
The valves were 12v solenoids from ali express, and the plumbing was from the hardware store. I almost guarantee it was far, far cheaper than this project.
Automation to determine that the thing watering the plants has failed is crucial.
This is especially true if the system is some form of setup in which there isn't a bunch of soil to buffer a couple of days of not watering correctly (like hydro- or aqua- ponics).
36 comments
[ 0.23 ms ] story [ 72.9 ms ] threadHome Assistant does a surprising amount of Disk I/O, if for nothing than for logs. Sibling commenters are also advising not running it on the SD card to avoid wearing it out, so there's definitely some truth here. This means we're adding a Pi M.2 hat + SSD into the mix. The Pi5 SSD kit for 256 GB, when it was available, was around $60 USD. A Pi5 with 8 GB of RAM is $130 USD. Now we need a cooler, a case that will fit the Pi5 with said M.2 hat, and a power supply. We're already well north of $250 USD, encroaching on $300, and we're not even using the core benefits of the Pi's platform. No need for GPIO pins, tightly integrated cameras or other sensors, none of that.
For all we know, the blog author did this assessment (or trusted the assessment of others, eg: Jeff) and came to the came conclusion - it wasn't worth the price of entry.
^1: https://www.jeffgeerling.com/blog/2025/intel-n100-better-val...
Since many devices are also Zigbee router (it's a mesh network after all), maybe some additional lightbulb for exterior would suffice ?
Docker compose with a zwave management server, reverse proxies for TLS, vlan isolation for the server, macvlan for HA container so it does see the host network, etc, etc. All to turn on and off a lightbulb with the sun. All the while AI is telling me to configure things insecurely.
I think when I get some more spare time, I’d like to write a statically linked program that handles a zwave controller and basic automation scripting. No IP networking needed for my lightbulbs. Then it wouldn’t feel risky to just make a system user and udev rule to give it permissions to the controller, and run with systemd.
I was trying to get some of the IoT I have at home like pool equipment, lights, HVAC, blinds, etc. Some of the setup were an uphill battle looking for more information in forums and trying to figure out what was broken.
Recently I decided over the weekend to use Claude and write a small app that controls my pool equipment and then deployed it using Cloudflare Zero Trust (kind of a reverse VPN). What a joy! Not only I had lots of fun reverse engineering my pool equipment API (I didn't want to depend on existing libraries - which I know exist) but I managed to create a fun and custom UI with React that my kids and wife love using. For example, whenever the pool heater is on, it adds an animated flame to the UI and change the background to a red-ish color. Plus it has a bar chart that shows the pool temp progression (takes hours to heat it up) with an animated volcano colors. The theme of the app is beach/pool vibe.
I don't think anyone here would be that excited if we were using the lower-denominator that HA turns out to be. I know it's a very cool automation tool, but just not very exciting and pretty obscure to configure every equipment I have at home.
I've been thinking about writing a blog post with the details of my fun project, let me know if anyone is interested in this. So far I've done the blinds and pool equipment. Next will be HVAC and lights. Took me 1-2 weeks total for each using Claude in my spare time.
I had the same thought after I joined a local group for Home Assistant users.
Everyone always talks about a happy path where you pick the right choices, use the right setup, and everything just works immediately. More often when people come to this local group's shared Slack channel it's because they're 10s of hours into trying to set up something that appeared to be simple. Then all of the old timers remember that they, too, suffered through something similar once and share what they can remember.
I think HA can be a lot of fun for people who like to experiment and debug, but if you're not the kind to be entertained by debugging your home's operation then it can feel like a chore. Some have an easy time setting it up and then get trapped when an upgrade breaks something or they try to add a new device with less than mature support.
I just bought Home Assistant Green: https://www.home-assistant.io/green and a z-wave dongle
That's about as easy as it gets.
I find home lab stuff has far more return on investment for like automatic blinds, lights, etc. It’s not like you can just stay inside anyways and get amazing vegetables, you still need to be on top of thing like checking for pests and disease. The automatic garden is a myth.
My home assistant also waters my plants. She generally enjoys working in the garden.
1) install HAOS, running it any other way is a step towards madness
2) get a decent zigbee controller
3) https://sonoff.tech/en-uk/products/sonoff-zigbee-smart-water...
Job done, the rest is all up to taste using either sensors or public data to infer water state.
Really makes me want to integrate the whole thing into an ESP32 with a display so such a thing doesn't require HA.
The valves were 12v solenoids from ali express, and the plumbing was from the hardware store. I almost guarantee it was far, far cheaper than this project.
Automation to determine that the thing watering the plants has failed is crucial.
This is especially true if the system is some form of setup in which there isn't a bunch of soil to buffer a couple of days of not watering correctly (like hydro- or aqua- ponics).