Problems downsampling data

#1

Hello - I’m a relative InfluxDB noob, using it to collect metrics from Prometheus servers for storage.

As we’re really starting to push our environments and getting more data in, we now have consumed about 200GB for 4 months of data. While we have plenty of storage, our requirements don’t need keeping that much raw data so I’m starting to implement RP’s and CQ’s to aggregate the data. So I created all new RP’s, shifted my default to the new RP, and now I want to copy data from the old RP into the new ones to free up storage. But when I run my downsample query, I always get:

SELECT MEAN() INTO “gkeedisontest”.“cq_5m_for_90d”.:MEASUREMENT FROM /./ WHERE time >= now() - 90d AND time <= now() - 30d group by time(5m)
ERR: no data received

This query run for a long time before erroring out. My series cardinality is 5.8M on this DB. Is it just too much for InfluxDB to handle moving this data? Ultimately we’re not very far into our use especially with out volatility in the environment, so I could just abandon the old RP after 30 days, but I would be interested in knowing why I can’t seem to get this query to complete. I’m sure I’m doing something wrong.

The server has 4vCPU’s and 64GB of memory.

Thanks!

#2

Hi ,
Your downsample query queries all measurements in your database ( /./) ,
You could limit the number of measurements in this query or make the date range smaller …

#3

There are hundred or thousands of measurements in the DB and if I’m going to move them, they all need to get transferred. I guess I could try moving 10 days at a time. But on the other side, I have 6 databases I need to do it to - that’s a lot of manual moving of data.