Not sure why my graphs look like this now

image

They used to be beautiful but after the last update they became like this. any idea?

Because this entity started being unavailable very often.

mmm that would be weird, I have a lot of entities that worked fine and all of them started being like this after the last update.

image

image

Check Settings → System → Logs for errors related to your recorder database or the sensors.

I don’t see anything like that in the logs

I’m keeping an eye on the logs but nothing seems related to the sensors. what else could it be?

This:

no, it can’t be, i have a lot of entities and all show the same pattern. I’ve noticed something though. today I had to stop a service that was making a lot of increments to an API endpoint and for a few hours it was actually not incrementing anything, after a while I’ve restarted it and it now all the entities seem to be consistent again. pretty weird.

What database are you running? The default SQLite or something else?

1 Like

it’s the default one. I’ll read how to implement mysql.
i’m almost sure it’s some throttling on the writes now.
I’ll let you know my findings.

I’ve installed mysql and configured it successfully, for a few hours it looked stable, and then this:

nothing weird on logs again. it’s really weird how it broke at 17:00

I’m testing this right now:



  - sensor:
      - name: "test"
        state: 1
        unit_of_measurement: "Mbps"
        state_class: measurement

image

this is not a sensor, it’s just a fake value, and the graph has a hole. i have no idea how to fix this now

There is smth wrong with database.
I would suggest an extreme way: stop HA, delete db-file, start HA & keep monitoring.
(+check hdd)

1 Like

but the database is brand new since i’ve switched to Mysql from Sqlite

the timing of the gaps make me think of a software glitch

i mean, this is too sharp to be a db issue:

Well if it is then it is in your system only. No one else is seeing this.

Use the System monitor integration to check your CPU and RAM resource use.

You could also enable debug logging and see if that shows anything up in the logs.

1 Like

ok, i will and post any result here. thanks

  • Entity sensor.ups_temp (<class ‘homeassistant.components.snmp.sensor.SnmpSensor’>) is using native unit of measurement ‘ºC’ which is not a valid unit for the device class (‘temperature’) it is using; expected one of [‘K’, ‘°F’, ‘°C’]; Please update your configuration if your entity is manually configured, otherwise create a bug report at Issues · home-assistant/core · GitHub

I’ve just started the logs and I see this, it looks like

ºC is different from °C (they are really different utf codes. ° and º)

Yes that’s a common mistake.

I’ve restarted the virtual machine and it’s almost 24 hours it’s recording stats just fine.

this is the static value to 1 in the last 2 days

image

I had the same issue - “messed” up my graphs without any changes I did.

I looked at my history and disabled “logarithmic scale”


Maybe this also helps?