• walden@sub.wetshaving.social
    link
    fedilink
    English
    arrow-up
    5
    ·
    edit-2
    1 year ago

    I took their recommendation and removed “purge_keep_days: xx” from my config. For the one thing that I’m interested in keeping data for an extended time, it defaulted to 10 days and deleted all of my old data.

    Looking into it further, I had to edit that particular sensor and add “state_class: xx” (measurement, total, or total_increasing) as described here to get long term statistics working.

    Time will tell, but hopefully I can start collecting more data. I’m bummed I lost my nice long term graph, but in the end maybe this will be better.

    Edit: Now I’m on a tear going through all of my custom sensors and making sure they have a state_class associated with them. I’m learning that in some cases, I can configure them through the UI as a helper instead of in YAML.

    • CmdrShepard@lemmy.one
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 year ago

      I’ve been using Grafana and InfluxDB to maintain device history as it can store everything for as long as you want. You might look into it if this new change doesn’t give you what you’re looking for.

      • walden@sub.wetshaving.social
        link
        fedilink
        English
        arrow-up
        2
        ·
        1 year ago

        Thanks, yeah I actually have InfluxDB set up and it contains history for the sensor that I “lost” so that’s nice. I don’t have Grafana set up, and InfluxDB seems to limit me on what I can graph. It says “Large response truncated to first 58k rows”.

        I’ll look into adding Grafana. Does it pull from InfluxDB?

        • eutampieri@feddit.it
          link
          fedilink
          English
          arrow-up
          3
          ·
          1 year ago

          Yes. And you should do something such as averaging if you want to fetch all the data I.e. to graph it. Grafana does it automatically iirc

    • thehatfox@lemmy.worldOP
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      It might be worth keeping purge_keep_days configured to less than the default 10 days to keep the database size small and speed up backups etc, especially in setups running on low power hosts like a Pi and with large amounts of sensors.