High CPU and Disk I/O usage

I’ve decided to remove unused measurements from my Home Assistant and also migrate 1 of them from an old name to a new name. Also added a couple of sensors for Wi-Fi RSSI data (updated once every 10-30seconds)
After that the Disk usage is pinned down to 50MB/sec write!
Even after disabling the Home Assistant export to InfluxDB (so no data is being ingressed whatsoever) the load is still the same.

Basically the load on the system did not change virtually at all (those added sensors are minesculem compared to all data already collected). Plus removing data ingress completely doesn’t help.
Tried uninstalling the InfluxDB add-on and restoring it from a back-up - that didn’t help.
The version is 1.8.10

Debug info available here influxdb - Google Drive

You can see that the average load before the breaking event is really low and completely handle-able by a Raspberry Pi with a fast SSD.

Hi @RiDEN,
So in a lot of cases, this is due InfluxDB running through reindexing and compaction of the new measurements. All be it the usage increase does seem significant. Could you provide a data example of the new measurements and the amount of tags introduced for each?

Strangely - I can’t remember exactly what have I migrated succesfully.
But now I can reproduce this behavior reliably by trying to migrate temperature data. It’s been collecting since April (each 10-20 seconds)
Here’s how it looks in the UI:


btw it probably duplicates the data. I mean it has a separate RSSI measurement, however the temperature measurement has an rssi field. a bit strange, if I understand it correctly.
So based on this - it’s only 1 tag (entity_id). If it’s not the case - please correct me)

Also it started randomly this night, and got fixed only after a second reboot. While after the first one the load got smaller :thinking:

By the way maybe there’s an easier way to rename the measurements ? I’ve got the climate data since way back, and don’t want to lose it when migrating to human-readable entity names in home assistant (and subsequently InfluxDB)

Now the issue appeared again, however this time it’s only 13MB/s.
Already rebooted like 5 times which didn’t help. Started doing this at a random time, while I was asleep 06:10 AM local time.
Any suggestions?

@Jay_Clifford hey, really sorry for bothering you, but I would be really grateful for any info/help/directions to look into and more than happy to provide any relevant information in order to debug this issue