Despite multiple restarts, recorder.purge services calls, only part of the data I excluded gets purged.
For instance, the data from my ADSL box is not purged - one example of an entity is sensor.neufbox_router_b_received even thoug it is listed as an exclude_glob expression in my configuration.yaml :
recorder:
commit_interval: 60
purge_keep_days: 14
# db_url: sqlite:////home/user/.homeassistant/test
# db_url: !secret ha_db_url
exclude:
domains:
#- device_tracker
- media_player
- uptime
- time_date
- worldclock
entity_globs:
- sensor.clock*
- sensor.date*
- sensor.glances*
- sensor.load_*m
- sensor.neufbox_*
- sensor.time*
- sensor.uptime*
- sensor.idfm*
- sensor.epson_*
- sensor.brother_*
- binary_sensor.idfm*
- weather.*
entities:
# - camera.front_door
- sensor.home_assistant_v2_db
- sensor.memory_free
- sensor.memory_use
- sensor.memory_use_percent
- sensor.processor_use
- weather.openweathermap
To undo any doubt about how I try to purge my database, this is the service call:
service: recorder.purge
data:
repack: true
apply_filter: true
I have two HA systems, and on the initial call history for quite a few entities did get truncated, which was noticable in the history view, but I lost only between 10 to 25 % of the data base sizes which still end up to be just over 400MB and 800MB on my systems.
Also, is there any tool other than the history view to help efficiently identify the entities that use up the most space? (There is a whole thread about optimizing the recorder)