Bug in rule machine?

So basically I have a rule for my sprinkles.
It sets the time to 1600sec but 1st zone only gets 50% of the time which means 13 mins.
When the rule triggers automatically everything seems to be working fine but because sometimes the weather has a false positive, I have to manually run the actions.
The bug seems to be happening only when manually triggering the rule, please see in screenshot the wait time before turning off the sprinkle is shown correctly (13 mins) but the sprinkle log shows being turned off only after 3 mins instead of 13 mins.
So I'm 99% this is a bug from rule machine since the time is shown correctly but not respected and since is happening only when running the action directly from the rule manually.
If it matters, most of these devices are meshed from a secondary hub and the rule is made on the main hub.

You'll probably need to turn on all logging for the rule - events, triggers, actions and post a screenshot from the logs when the problem occurs.

I enabled them and continue to monitor the problem in the next days but where are these logs shown more exactly?In app status?
Some dev should probably try to reproduce the problem and see if it happens to them too.

No. At the bottom of the left hand column while looking at the hubs page in a browser. Just click on 'logs' and then 'past logs'. You can then select your rule in the filter - click on the link at the top of the page. You have shown the events from the device and that will show nowhere near enough detail. The logs from the time of the problem will show everything that occurred in order so that it's possible to see what did (or didn't) happen and why. However you need to check in the rule itself that all three boxes for logging is enabled - triggers, events and actions.

If you can easily make this fail by running the rule manually and logging is enabled already, there's no need to monitor it for days. Just manually run the rule now and once the actions have completed, post a screenshot of the logs for the rule (starting with the manual trigger)

No one will want to recreate your rule and attempt to find an issue that could be an error in your rule logic. You first need to prove the issue by providing the detailed logs. You might still have the required information in 'past logs'

1 Like

I just ran again the actions manually without changing anything from the rule and now it worked as expected, very strange, if is a random bug is even worse but is clear that it happened the 2nd time in a week (that i'm aware of ) .
Will keep the logs enabled and post them here when it happens again.

Please show this rule in its entirety. Including triggers and all the variables. The portion you’ve shown seems to indicate a flaw in the rule’s logic.

Logs would also be useful.

1 Like

Note that I reverted the rule to main time 800sec that this is the original rule.

Yes, do that. Those logs will likely show what went wrong/what wasn't actioned. Also post what firmware version your hub is running, as Rule Machine bugs that are reported and confirmed tend to get fixed by Bruce in the firmware updates. Current version is

1 Like I'm having the last version, will keep you posted when it happens again, thanks

1 Like

Anyone else think this? Maybe I've been doing keto too long...


Are you sure it's not Rule Machine working as designed? The delay you circled is not marked as "cancelable" (but your others are), so if your rule triggers twice in the specified time, the "old" delayed action will still be there even after the new one gets scheduled. (This includes "manually" running the rule from Run Rule Actions.) Thus it may, as you described, appear to turn off sooner than expected -- but that's why.

It should also be noted that even with this option enabled, the delay would actually need to be cancelled somehow (a trigger alone doesn't do that). I do see one place in your rule where that could happen. If your other delays are working as expected, that is almost certainly the problem with this one.

If not, I second the above: enabling logging will be your best bet for troubleshooting. The docs tell you both how to do this and where to look for them.


wow, eagle eye .
I replaced one zone with another some weeks ago so I had to replace them in the rule machine too, somehow I forgot to make it cancelable and now I missed this because I'm 100% 1st time I created the rule triple checked this aspect.
Not even need to test now, is clear to me that this was the problem especially since the hub was not responding so I ran the action several times and this caused the problem. Thanks a lot.