Application takes a lot of hub cycles

[vitaliy_kh]

I have one really simple application which takes a lot of hub resource.

Here is an application in question:

Basically it runs on a periodic schedule (every 12 sec)
and sends a message to the PixelBlaze Driver.
I have about 7 PixelBlaze Controllers but only this one requires
periodic communication.
What could be wrong with my RM rule or PixelBlaze custom driver?

@gopher.ny
UPDATE
According to the Appstat page this RM Rule in question takes more than 50%
of hub resource. This means every minute more than 30 sec hub
is busy with processing this RM rule. Rule is on a 12 sec periodic schedule.
If this statistic is correct hub has no choice to run anything else.
However hub runs everything I would say OK.
BTW,
once I saw "% of total" column was showing well above 200% value
(unfortunately I missed to catch this).
Question is:
How trusty all this statistical data is?
I think it could a bug in the reporting algorithm.

[bertabcd1234]

12 seconds is a very frequent schedule. I'd say it would be normal for that to be a large percentage of your hub time. What we can't see from your screenshot is if the "more than 50%" that you state is percent of total or percent of busy. That's higher than I've seen for anything (that I've looked at) for percent of total, but as a percent of busy, it could still be quite low -- keep in mind that the percent of total would then be a percent of that percent, ultimately a (usually much) smaller number.

If you aren't actually having problems, I suppose there's no reason to be concerned; these are there as one way to help you identify possible causes if you do. I'd avoid frequent schedules unless you really have a need (because responding to events is generally a better approach), but if you do have such a need, then I don't know what else to suggest.

@vitaliy_kh

once I saw "% of total" column was showing well above 200% value
[...]
I think it could a bug in the reporting algorithm.

No, the total is per core. Hubitat uses a quad-core processor. Therefore, 200% is in range (though probably not something you should see often).

[vitaliy_kh]

@bertabcd1234

What we can't see from your screenshot is if the "more than 50%" that you state is percent of total or percent of busy.

Here is an updated statistics:

@bertabcd1234

If you aren't actually having problems, I suppose there's no reason to be concerned;

@bertabcd1234

I'd avoid frequent schedules unless you really have a need (because responding to events is generally a better approach),

This RM rule was running on C7 more than a year. Frankly I did not look on a statistic page since C7 was running very well (unless I screwed up something myself). After migration to C8 (which went very well) I started to see High Load warnings well too often and statistic pointed to this specific rule. BTW, the original rule had 7 triggers (one for each parameter change). Since I started to see this High Load warnings I changed trigger to periodic 12 sec scheduled run. High Load warning disappeared but statistic for this RM is still showing very high load (unless I am misunderstanding something). So now I am confused, do i have a problem and if "yes" how to fix it.
Any ideas how to fix a problem (if this is a problem) are more than very welcome.

[bravenel]

@vitaliy_kh

So now I am confused, do i have a problem

Not if the rule is working and your hub is working as you expect otherwise.

[vitaliy_kh]

@bravenel

Not if the rule is working and your hub is working as you expect otherwise.

The original RM rule with 7 triggers for each variable change was running on C7 for more than a year and there was not any High Load warnings never ever. Because of this I never looked on a statistic page on C7 so, I have no idea what these numbers was. After migrating to the C8 this specific RM rule produced very frequent High Load warnings. I changed triggering to a Periodic Schedule every 12 sec. But according to @bertabcd1234 message this is too often and not recommended. With Periodic Schedule trigger High Load warnings disappeared but statistic for this specific rule is still too high (I think).
And as result I am confused should I keep Periodic Trigger or should I change it back to multiple triggers for each parameter change and simply ignore High Load warnings.
Please advice.

[bravenel]

This post is virtually identical to your post just above.

What was wrong with 7 triggers? That sounds better than periodic. How frequently did it trigger?

[vitaliy_kh]

@bravenel

What was wrong with 7 triggers? That sounds better than periodic. How frequently did it trigger?

Absolutely nothing was wrong on C7
C8 started to scream about High Load.
There are 7 weather related parameters from local Ecowitt Sensors. Once in a while (depend on weather conditions) changes could be quite often but I don't really know how often they are.

Another difference in my setup is - now Ecowitt Sensors integration is on C7 and sends all data to C8 via Hub Mesh.