Rule condition shows TRUE in UI but Rule never transitions to SET — survives full restart

[toggledbits]

Does that sensor reliably come up "off" when you restart HA?

[CatmanV2]

I don't believe so, It's been used in several places in HA without issue across restarts. I will see if I can double check that however, as it may simply not have been noticed if it's coming up 'off' and immediately tripping to the correct state.

C

[CatmanV2]

OK I've just done some digging: When HA restarts, all three sensors

sensor.rachel_primary_region

Go 'unknown'

Someone Home (fairly logically) goes 'Off'

The period of 'Unknown' is about 5-10 seconds before the correct status is realised.

Here's the data from the last restart

binary_sensor.someone_home	unknown	2026-05-19T12:06:55.461Z
binary_sensor.someone_home	off	2026-05-19T12:07:04.582Z
binary_sensor.someone_home	on	2026-05-19T12:07:11.505Z

TIA

C

[therealdb]

It may. I have similar behavior with other Template sensors being 0/null/off at boot.

As a defensive measure, I always make the condition to be sustained for at least 10 secs and/or avoid them being evaluated when ha has been restarted for the last 20 secs.

I rarely use ha templates, so my experience is anecdotal.

[CatmanV2]

Thanks, helpful input!

C

[therealdb]

I think you may just filter for unknown and return null. Tell that to my friend Claude

[CatmanV2]

@therealdb that would rather force (in this case) all the heating off (or on), wouldn't it?

I like the idea of a 30 second delay though.

C

[therealdb]

I don’t know if setting to null it’s better for reactor. @toggledbits here is the best source of truth.

[CatmanV2]

Of course. I'm not sure what I'd do without him!

C

[toggledbits]

OK. I can't prevent HA from starting up in an undefined state, and the effect that can have on the rules, but @therealdb is spot-on with an approach to dealing with that, since the corrected state usually comes pretty quickly after. Sometimes, you just have to debounce those switches a little.

Beyond that, there is definitely an opportunity for a race condition and I think I have a good general fix for that. The logs support that cause as well. Testing my updates now. Will likely do a build with the update tomorrow.

[CatmanV2]

Fabulous!

I've actually just put reset delays in the two rules that are triggered. The 'set' is heating on so it's the reset that's a PITA. That should cover it.

C

[toggledbits]

This is just a follow-up to link this conversation to the announcement for the build that addresses it, and expands generally on transient startup/outage states in HA integrations

[CatmanV2]

Thanks, installed and working.

Although after the first restart there were some virtual switches in HA that were not being accurately reflected in MSR, but I restarted MSR again, and everything looks fine.

Thanks again for your help. Next time I'll upload the logs to start with!

C