I have a lot of lights, about 100. My biggest group has about 25 lights.
I automated all my lights individually (auto brightness adjustment based on outdoor sensors lumens, colors defined for each moment of the day and so on).
I noticed that some requests are lost when I send a lot of requests at the same time.
I found a workaround adding a 200ms delay between queries and it looks like reliable now.
My question is quite simple: it is possible to improve queries management to avoid this use of workaround on client side?
Don’t you get a bridge busy message if you do too many requests? In pydeconz I schedule retries if I receive the bridge busy error message deconz/api.py at 60a07ba43f372fd28085199bfb841acce4fafb1c · Kane610/deconz · GitHub
Thanks, it is a better approach than mine. Busy is probably 503 HTTP response.
I send API requests using node-red (http request) and using PHP (curl).
I have to implement this queue & schedule management in both side. You implemented it using python.
Is there no way at all that deconz increase simultaneous http connections or handle this queue itself?