This isn’t an access issue. Your HA core container is crashing.
The fact that it returns to normal after a restart seems to indicate something may be using up all your RAM or CPU. You can monitor this with the System Monitor integration.
If you use loops in your automations or scripts check these can actually exit. Infinite loops eat RAM.
It could also be due to a rouge 3rd party integration. There is some advice here on how to track that down: 2024.5+: Tracking down instability issues caused by integrations