On my Homey Pro early 2023 When-Cards of type. “Every n Seconds” (e.g. Every 1 Seconds) randomly (e.g. after months) and unpredictably stop working in various (not always the same) advanced flows.
After creating a new identical card and replacing the existing one with it events immediately fire every second again.
I can not relate this to any unusual events like updates, backup restores, power outages etc. The flows in which it happens had not been touched for weeks or even months.
Did anybody experience the same?
What did you do against it?
In overload situations Homey usually does not disable single cards but complete flows and marks these as disabled. I am also monitoring CPU load and there was nothing extraordinary.
Very likely that the flow gets rate-limited for being called too often. This is unrelated to CPU load, it’s just a dumb “when the flow is started X times within Y amount of time, disable it” method.
@4x4_Pete
Additionally, there’s a “when a flow has more than 60 instances running, it will be disabled” (without warning) limit.
Meaning, when a flow finishes in 2 minutes, due to timers/delays, after one minute it will have 60 instances running when triggered every second.
Question: why do you use / need “every second” triggers?
Hi @4x4_Pete , Every n seconds is very often, and much more, all of them start at 00:00 midnight every day, as well as every n minutes/hours.
But to identify any other reason, even if it looks randomly, we would need more information, like how many are they, which are those which stops extraordinarily often. Do they recover without any actions from your side, after what time, etc.
Probably one of us finds one or more additional reasons.
Regards Dirk
You’re right and important for a request to Athom, if that is the problem.
Never the less in every machine is a limit. So for me as user of an existing system it is rather important to detect the limit and to learn how to avoid them.
And even if I have with my few time dependent WHEN cards, I use Cronjobs and distribute the different flows to other seconds.
True, but even if your Homey is doing absolutely nothing, flows still get rate limited, because it’s not based on system load but on hard-coded values.
My “Every n Seconds“ flows usually just do simple computations that finish within some milliseconds.
But the annoying thing is this:
When flows exceed some quota by e.g. getting called to often, the complete flow gets deactivated, which is then clearly visible and easy to detect.
In my case, however, only that single flow cardwithin the flow gets somehow deactivated while the flow itself remains active, which is very hard to detect.
Hi @4x4_Pete , @robertklep is right, you should contact Athom support on this.
Nevertheless I would propose to avoid these triggers where possible or distribute them as much as possible via Cronjobs . If you don’t have an idea, you could share some of the flows and some of us might give few ideas. Then you can decide, if you want to use them.
Kind regards
I had filed the case with Athom already and they vaguely admitted that it might be a bug indeed. Additionally posted the case here to check, if that problem is more common and if someone has successfully solved it.
Unfortunately it so far seems that I am more or less the only one experiencing this weird glitch.
I don’t think that’s an actual function or some sort of security feature.
You indeed probably found a bug (like Athom ‘vaguely admitted’), if the flow itself didn’t get deactivated.
