You should be able to continue at the point it crashed when you have half CSV.
The measure tool should provide you with an option to resume.

Will have a look how to improve retry mechanism when HA API is temporarily unavailable.

By default the RESUME var is not set in the env. I uncommented it out and resume worked. However still having timeouts so having to keep restarting the script.

Ok so I figured out why I'm getting these random 500 errors.

I checked the Home Assistant main log and see.

Logger: homeassistant.components.light
Source: components/light/__init__.py:331
integration: Light ([documentation](https://www.home-assistant.io/integrations/light), [issues](https://github.com/home-assistant/core/issues?q=is%3Aissue+is%3Aopen+label%3A%22integration%3A+light%22))
First occurred: 13:40:12 (5 occurrences)
Last logged: 23:38:17

Got `kelvin` argument in `turn_on` service, which is deprecated and will break in Home Assistant 2026.1, please use `color_temp_kelvin` argument

Actually it looks like the issue is really the Hubspace integration. I keep getting 403 from their web service, my guess is rate limiting maybe occurring.

Yes the "error" you showed from log is not really an error but just a deprecation notice. Shouldn't break anything, and not be the cause of 500 error.

Actually it looks like the issue is really the Hubspace integration. I keep getting 403 from their web service, my guess is rate limiting maybe occurring.

You could try raising the SLEEP_TIME in order to prevent hitting rate limiting on that cloud API. looking at the response code it shouldn't be rate limiting, but I won't be surprised if it's a badly implemented API or infrastructure.

Well majority of my issues were during the color_temp. Once I got into the HsVaration all my 403 errors resolved.

I have probably fixed the deprecation error with #3041
Might be the problem you were running in the 403 errors, as the hs didn't give you issues.
Good to have this warning resolved anyway.

However having a hard time building the docker image.
github action keeps running in strange segmentation faults after 20 minutes :-(
New run is running here: https://github.com/bramstroker/homeassistant-powercalc/actions/runs/13077153943
Also cleared github caches now. Hopefully will succeed now.

Yeah thats what I'm thinking it was Hubspace really causing the issues by sending a 403 randomly. And Home Assistant just returning 500 instead of the actual message. The other thing was just having to restart it after every 500 that was the real pain because I would start the measure script. Go about my day and come back to find out it crashed 5 minutes after I started it and now I gotta resume it. So it was really having to babysit the measure script to ensure it was still working.

Closing this issue. Improvements have been done in measure tool to prevent the script from crashing on one error.

Yep seems to be working great been running the effect measure for over 30 hours and no unexpected quits.