Databases sizes - Recorder and InfluxDB

Configuration
1

I am having an issue with database sizes. I’ve switched my recorder over to MariaDB and I’m also running InfluxDB. Problem is, despite attempting to limit sizes, they are getting huge! Here are my settings:

For recorder:

recorder:
  db_url: mysql://UID:PASS@core-mariadb/homeassistant
  purge_keep_days: 1
  purge_interval: 1
  exclude:
    domains:
     - updater
     - automation
     - calendar
     - camera
     - media_player
     - remote
     - zwave
     
    entities:
     - sun.sun
     - sensor.last_boot
     - sensor.date
     - sensor.moon
     - sensor.next_launch
     - sensor.prusa_mk3_actual_bed_temp
     - sensor.prusa_mk3_actual_tool0_temp
     - sensor.prusa_mk3_job_percentage
     - sensor.prusa_mk3_target_bed_temp
     - sensor.prusa_mk3_target_tool0_temp
     - sensor.prusa_mk3_time_elapsed
     - sensor.prusa_mk3_time_remaining
     - sensor.yr_symbol
     - weather.dark_sky

For InfluxDB:

i nfluxdb:
host: a0d7b954-influxdb
port: 8086
database: homeassistant
username: !secret influxdbuid
password: !secret influxdbpass
max_retries: 5
default_measurement: state
include:
domains:
- sensor
- binary_sensor

Additionally, the retention policy in InfluxDB is set to 7 days. There are about 90 series showing in InfluxDB for my various sensors, most of which I kind of want. I can probably exclude a handful more.

Yesterday, I dumped InfluxDB and re-initalized it, just to clear old stuff. MariaDB was dumped a few days ago. Today I took a snapshot, MariaDB is taking up 186 MB, InfluxDB 358MB.

Why are they so huge? Is there really that much data in there, or am I doing something wrong in my settings? I was hoping to use InfluxDB to track months worth of sensor data, expecting it to store them quite compactly, but at this rate it’s not really feasible if a day’s worth is a 1/3 of a GB. For the recorder, I mostly want it there so that state is preserved during reboots.

Any solutions other than to just filter entities more tightly?

Thanks.

2

my setting is purge_keep_days: 2

# hass-cli state list|wc -l
322
# du -hs /var/lib/mysql/hass
447M    /var/lib/mysql/hass
#
# du -hs /var/lib/influxdb/
497M    /var/lib/influxdb/

I don’t think i’m purging influx. Look inside, see whats in there.