It just slightly annoys me how the rage is all around how the evil corporation locked us out, because they hate their customers, etc etc, while conveniently ignoring the fact that HA has been hammering their API for years to the limit of what could absolutely be considered a ddos and they tolerated that until now. Why now ? Who knows. Maybe it’s indeed a business strategy thing because they want you to see their ads. Maybe they were unhappy with the 6k-something HA MyQ users sucking up so much API processing time, making the experience more laggy for their other millions of customers. Or maybe they just had an internal cost cutting round and the server costs showed up on the balance sheet. There are always two sides to a story like that.
Of course their way of reacting by completely blocking access wasn’t ideal. They could have reached out to the HA devs, proposed a partnership with proper access to their API (with push functionality instead of polling). Or even better, they could have opened the API for everybody (although that comes with a lot of legal and financial caveats). But they didn’t, and that’s their prerogative. It’s their API, they pay for its maintenance after all. The fact that the HA integration massively abused their API for years probably didn’t help. High frequency polling a cloud API is never a good idea. Ever.
Honestly that should be a minimal load. How does their app do it? How often does it poll? You have to compare the load from the app vs ha to make any valid assessmemt of whether it can be classed as spamming their servers.
That’s true. But, as far as I know, their official app doesn’t poll at all. Polling is considered really bad practice in modern webapp design, because it puts so much stress on the servers for essentially nothing. Apps will now normally use a push scenario, where the server tells the client app when something happened. That reduces the amount of stress put on the servers and the network by orders of magnitude. So if a server is designed and sized to primarily handle push requests, it can very easily be overwhelmed by high frequency polls.
Yes that would have been a good way to handle the situation. I think the HA MyQ integration guy actually asked them for a partnership but they declined / did not respond. Maybe this is something a corporate entity like NC should initiate.
According to the developer working on the myq integration the official app uses mqtt somehow and doesn’t poll. He makes a lot of good points on this post about how this whole mess could have been avoided.
And on that note, I’ve definitely pulled mine off line. Wouldn’t be the first time a firmware update ended local control
They don’t have a web app, just mobile apps. My guess is they rely completely on iOS / Android push notifications, which HA can’t hook into.
Another way to look at it is at those numbers, the integration would be responsible for ~218 requests per second. That’s not that many… but if they didn’t design their services to expect that, that could easily hit a slow path. Of course, they could throttle it without completely blocking, or make their service more performant, or even consider charging users a monthly fee for API access.
What’s interesting is between this and the Mazda integration, there’s been broader press coverage of the existence of these sorts of cloud integrations. What happens to adoption of systems like Home Assistant if the majority of cloud polling / push integrations get blocked over the long run? It seems like a real threat to the project.
It makes sense the bridges use mqtt to publish states and listen to commands, but I doubt they listen to states of other garage doors. Even if the app uses mqtt when open, it can’t maintain a persistent connection for more than a few minutes once it’s closed.
Regardless, based on Chamberlain’s past history of trying to use the DMCA to block universal garage door remotes, it’s clear they’re not a good actor in this space.
From what I can see, The Chamberlain Group isn’t interested in any sort of home automation server integration. They are seeking multi-year subscriptions tied to each vehicle app as their angle: Connected Car Services & Solutions | myQ
I considered a Meross MSG100HK (with the free yellow button adapter) connected to Home Assistant. But instead I got a Security+ 2.0 Dry Contact adapter from EBay (basically a modified 883LM button) wired to a WiFi relay and I’m hoping to integrate that from Home Assistant to CarPlay.
I’m leaving the MyQ there to see what happens with Amazon delivery and the car apps I use (which aren’t supported yet), but I’m not excited about having to pay for a 3-year or 5-year subscription to use MyQ in my car. I’d love to be able to talk from my Home Assistant to the MyQ device via mqtt, but I don’t think that’s going to happen.
Can you give some more info on this? What opener did you have - Security +1 vs 2? What kind of MQTT commands? I know that the door buttons at my house simply put different value capacitors across the two wires for the lock and light buttons, and I could duplicate that if I had to, but if at all possible, I would like to just be able to do it directly.
My Homebridge hasn’t connected to the internet since the day I set it up, so I could care less if I loose access to the cloud. My concern was with the removal of the MyQ integration.
I’m hoping since I’m using HomeKit i’m not affected and this is only for the cloud connection.
I’ll get RatGDOs at some point but so far homebridge has been okay and was easy to setup.
I’m unaware of any way to get video outside of myqs apps for their video products- their system by design does not play well with other options. They want you to have to buy their video subscription and not be able to use the camera with anything else.
I would recommend installing a generic ip camera nearby that uses a universal protocol like mjpeg or onvif that will easily integrate with home assistant
I have a foscam g2 on my garage but see this post for some other camera suggestions
I ended up using the Konnected garage door control. I also purchased a Security 2.0 compatible button from Amazon for $11. Toss in a little soldering and I have a complete, local control solution to replace the crappy MyQ. The Konnected box has a door open sensor that measures the distance from the ceiling so it can tell when the door is open. It’s not as well instrumented as MyQ was but it works.
It’s also running esphome so you can customize it if you want. Highly recommended
And when it comes to IP camera, outside of the typical options of a dedicated camera in the garage, there is also the option of installing IP Webcam Pro onto almost all unused Android phone, if you have any in your junk drawer.
Over the years I have gathered 3 old Android phones from my family. So I paid for the Pro version of the app, which would turn all my unused Android phones as regular IP cams.
The problem is the communication beween the opener and the gateway is secure and proprietary, so you are unable to acess it or control it.
Yeah, I know that I need something else, and I’m planning on removing the MyQ gateway, since the company seems insistent on basically making it useless to me.
So I’m thinking I’ll go with a ratgdo also, but I’m trying to figure out what the likelihood is of success in getting the lockout to work on the remotes, since I have the v1, and the docs say it doesn’t work, but you seemed to say that you got it to. So I’m trying to mainly get more info on what you did before moving ahead with this. I want to do my homework, and if I need to build a custom board to support the “Lock” feature then I will do that.