Influxdb 1.7 - high ram - No queries

1

Heya All, I’m at the end of my rope and I thought before I call it a day and try again tomorrow I would ask here.

InfluxDB lives in a docker and it uses almost all of its hosts 32GB of ram, TSI enabled. Our biggest (by far) database within influxdb has a cardinality of 4065089.

Nothing much has changed in regards to queries or load on the server and it normally has a fairly big resource footprint. Normally it runs at about 16GB stable and if we do a huge query, it might shoot up to 2x’s and then reduce.

I ran a test to see where the load was coming from by removing the database from grafana. However even after completely stopping grafana and rebooting the influxdb docker, the ram began to rise again, back up to 31GB with no load (grafana still stopped)… This doesn’t make sense to me… lol.If anyone can help, or point me in the right direction that would be great. I’m a novice when it comes to influxdb so any help or pointers would be greatly appreciated.

2

Have a look at the influxDB logs and compare the timestamps (which are in UTC) with the high memory usage period
From time to time InfluxDB compacts its files, and it takes memory to do so.
I run on Windows and had some Out Of Memory errors during file compaction in the past.

For more detail about what happens in Influx you can have a look at the Storage Engine Concepts

If you don’t find anything useful you can try to change the logging level to a more verbose one (this requires a restart of Influx)

3

Heya Giovanni, thanks for your answer :slight_smile: It’s always high, like, there isn’t anything (apart from a big query) which changes it.

We were thinking that maybe it has something to do with the cache size and if reducing that, or turning it off completely will improve things. I know it’ll result in longer load times, but maybe we can always increase it to see if we can find a sweet spot.

series-id-set-cache-size