I’d recommend using this script:
To fix your Flows, if you still have the old device ID at least.
I have no flows with Tuya, i operate Tuya only manually with Desktop PC via webportal.
Flows from other App companys tested, sometimes work, sometime not, need restart Company app für Homey Connect… few options… That’s here ot… Not all Tuya devics working with Homey. Alltimes back to smartphone. But want control all devices by PC with big screen anmd alltimes on. 
In Germany we saying “Nichts Halbes, nichts Ganzes.”
After app restart, the door sensor via the Tuya Zigbee hub works again.
FYI
Sadly it still doenst work stable.
I constantly have to restart app to have connection with the devices. I automatically restart it every hour with a flow, but thats not enough apparently.
With Smartwares app devices always work instantly, so it has to do with Homey for sure.
Same issue here
At this time we are pretty confident that any issues that remain are solely caused by Tuya still enforcing an API limit that is a bit too strict. We just asked them again to relax this limit, which is unfortunately all we can do at this time.
After the latest update, it has been very stable at my home over the past few days. But I only have 4 Tuya lights and I control them twice a day with a flow (turning them on in the evening and turning them off when I go to bed). Maybe my configuration isn’t reaching the API rate limit.
It might be helpful if it doesn’t work for you, that you indicate how many Tuya devices you are using and how many calls you make in flows, because that way we might find a common thread (read, API rate limit).
What I mean: if a flow polls for information, you get more API calls than if I turn on 2 LED lights in the garden once it gets dark.
The rate limit only concerns the token endpoint, so the amount of devices is not a factor here.
Assuming that Tuya is rate-limiting on client ID, and the Homey app is using the Home Assistant client ID, does this mean that these issues are now also causing problems for HA users?
Probably rate limiting on IP. Else, a spammer/DDoS can just spam requests there from 1 IP and get the entire API taken down for everyone. That would be strange
But does the Homey app generate that much traffic that single users are being rate-limited?
Yes, it is affecting Home Assistant users as well.
Endpoints and domains are different things.
No, it is a generic endpoint rate limit that is being hit.
I know, but is it the endpoint on m1.tuyaeu.com?
That’s strange. So is it per endpoint per IP, or just per endpoint in general? If it’s a general rate limit, that would mean that anyone can take down the API by just spamming a few hundred requests to it, which would break the API for everyone.
I share the ID with my HA installation and when Tuya failed on Homey, on HA it just worked. 
The API endpoint actually differs for each account, so it could very well be that that is the one used for your API calls, yes.
You usually have multiple rate limits which can depend on multiple factors. This can be a relative low failure rate limit, a much higher generic rate limit on client id, and another one specific for an IP. But we are obviously not aware of the specifics at Tuya, just that the general limit was reached that also affected Home Assistant users.
That is very well possible. Not all users have the issues as it is all timing dependent (which is also good for us, as with the amount of users with this app answering all support request would have become impossible).
Sorry, I cant understand what the problem is with endpoints and rate limits etc.
What can I do to improve stability?
