The recorder backlog queue reached the maximum size

Since 5th July, I’m getting the error in the title happening on a daily basis.

The full error is:-

2024-07-09 09:38:56.637 ERROR (MainThread) [homeassistant.components.recorder.core] The recorder backlog queue reached the maximum size of 178958 events; usually, the system is CPU bound, I/O bound, or the database is corrupt due to a disk problem; The recorder will stop recording events to avoid running out of memory.

This seems to be related to a CPU spike that happens at 4am (When the CPU goes to 50%) every morning:-

I ran the command:-

sqlite3 home-assistant_v2.db 'pragma integrity_check;'

Which returned ‘ok’.

I don’t have any automations that trigger at 4am, and the dips in the chart above are me restarting HA.

Any clues on how to diagnose this ?

I’m running HA as a VM on proxmox with an SSD.and its had no problems in a couple of years.

FYI, I did update the system to the current version (2024.7.1) on the 4th July… so there might be something in that ?

1 Like

Pretty sure mine is ISSUE 1. The CPU spikes at 4.15am every morning, but the recorder does seem to hang on for a while before it finally fills up its queue.

I’ll try the auto purge: false and see what happens tonight.

1 Like

see my comment: PSA: 2024.7 recorder problems - #57 by zigomatic

Just to confirm, pausing the auto purge has solved my issue, so i guess I’ll leave it like that until release 2024.8 is released.

1 Like

Fix is out with 2024.7.2. :slight_smile: Backup and then update, should solve the problem.

See here (and some posts above this one):

:+1:

My recorder is still crashing and I’m on 2024.7.2. Is it just me? Going to try reverting to v2024.6.4.