Request help for Influx statistics report on data consummation like top 10 in size and quantity

Hi,
New here but been using Influxdb for a few years with my HomeAssistant project. The DB has growth to 1.6 GB and I would like to explore the data in statistics fashion where I can see the top offender in size or quantity to adjust the retention.

I’m trying to get familiar with InfluxQL and Flux with “Explore section” but I can’t seem to find a way to extract such reports.

  • Top 10 size bytes per Measurement, Tags or Fields.
  • Top 10 quantity count per Measurement, Tags or Fields.

Any other stats that can help me understand where this space has been consumed or tools to use is welcome.

Thanks

Hi @starwolf73,

Can you let us know what version of InfluxDB you are using?

I’m using Influxdb 1.8.2 with Chronograf 1.8.6.

Look into the influx_inspect report command, that might give you enough data to identify your top offenders.

Thanks @mhall119,
So I went over the documentation for Influx_inspect and finally was able to get something out to point me in the right direction but still far from what I was looking.

Using a combination of linux command

  • du -sk *|sort -n (Find the biggest TSM files)
  • influx_inspect report -detailed (Provide a narrow view inside this TSM file only)

The documentation on report command is missing some mandatory information like this was pointing out in this github issue. docs: add detailed info how to use influx_inspect to check if I use TSM or TSI · Issue #1703 · influxdata/docs.influxdata.com-ARCHIVE · GitHub

When using influx_inspect report command, the documentation is missing the following:

Report summary

  • influx_inspect report <path of the DB>

Details summary of a TSM

  • influx_inspect report -detailed <path of the TSM files>

So after compiling all those informations and using Explorer data GUI, I was able to start discovering some offenders but I’m far from a Top 10 measurements table like I use to do in a simple SQL DB query.

I’m new to this concept of shard and still need to digest more if there is other method to get a similar report.

Any other suggestion will be welcome, thanks.