That code is card config, so you add a card, switch to YAML, and then add the code.
Yes, in the meantime, I discovered more detailed instructions at How-To Geek, but the YAML listed there gave errors - I think that some indentations are incorrect. However, the YAML listed at github works just fine …
Now that I have it all working, I just have to interpret the results and improve my HA rating … 
This is a fun tool, thank you!
I’m running HAOS in a VM on a laptop with 4 CPUs, 6GB of memory, and a 32GB partition allocated for the VM. For some reason my score is stuck at 86%. When I look at the hardware items they are all below the limits you show in the README (10%, 70%, and 80%). I don’t have any zombies, my system is fully up-to-date, my database is 66MB, and I backup daily (and the backups are working). Here’s what my dashboard looks like:
It seems like my score should be 100%. Am I missing something?
Yes, it is a fun tool and has already helped me fix some health issues. But, I have 42 zombies (!) and this is my rating:
I would assume that my rating should be lower than yours given the 42 zombies detected?
I am not concerned about the zombies as I know what they are and they relate to “work in progress”.
Hah, I figured it out. I had enabled both sensor.system_monitor_memory_usage AND sensor.system_monitor_memory_use. It was grabbing the latter, which rather than being a percentage is the actual amount of memory used. Once I disabled the “_use” entity my score is now 100%!
hi there, thank you for the feedback. i have a v2.2 list that i will expand with some of your suggestions, thanks!
first off, shrinking a db from 34gb to 72mb is absolutely wild.
the reason you are stuck at 86% with a clean sheet is likely a mix of two suspects. first, double check your disk sensor. if you accidentally fed HAGHS a percentage sensor instead of absolute gigabytes, it might read something like “14” as 14gb and silently judge you.
since you mentioned ignoring items, you almost certainly hit the ghost penalty. in the current version, ignored updates or entities disappear from the card but silently keep deducting points in the background. we are actually in the middle of testing the v2.2 update, which completely fixes these ghost penalties and brings a foolproof ui config flow for the sensors. for now, just wear that 86% as a badge of honor.
oh wow, happy optimizing!
that ghost zombie is hiding in the shadows of your sensor attributes. just head over to developer tools → states, search for your haghs score sensor, and check the attributes column. the exact entity id of that single zombie will be listed right there. once you have the name, you can hunt it down and delete it. we are actually rebuilding the dashboard card template for the upcoming v2.2 update so these ghosts won’t be able to hide from the ui anymore.
thanks for upgrading! i would recommend you to use the new dashboard card examples in the github repo, there is a lite version and a pro version. this should solve the problem: GitHub - D-N91/home-assistant-global-health-score: An architectural framework for measuring system health and hygiene in Home Assistant. · GitHub
Thanks @denzo - I installed the “pro version” and that fixed it … 
glad i could help you out 
UGH…says I have 345 zombie entities. Is there an easy way to clear them out? Or do I have to do it manually?
345 is a lot, but there is a built-in bulk tool for this:
Go to Settings > Devices & Services > Entities, then filter by status “Unavailable”. You can select all of them and delete in one go.
Before you do that, worth asking yourself two things:
- Are these devices actually gone? (old integrations, removed hardware) → delete them
- Are some of these intentionally offline? (seasonal devices, tablets that sleep, etc.) → use the
haghs_ignorelabel instead of deleting. Assign it to the device and HAGHS will exclude all its entities from the zombie count permanently.
The HAGHS Pro card also shows zombies grouped by domain, which makes it easier to spot patterns before deciding what to do.
