I am trying to display with grafana 2 different retention policies on the same dashboard. (one retention policies contains long term data downsampled)
It is working well on this kind of query : select mean(/(mean_)?idle$/) from "$RP"."system.cpu.total"
depending on the retention policy RP, it displays either “idle” or “mean_idle”
But if I want to display the cpu “not idle” I can’t … select 100 - mean(/(mean_)?idle$/) from "$RP"."system.cpu.total"
This query returns me no data points for any retention policies …
Is there a limitation on influxdb side that forbid me of doing this ?
Should I open an issue on github ?
downsampled_data_5m (data from RP raw_data aggregates into this RP every 5 minutes).
At Grafana dashboard I have variable $retention_policy, that equals to raw_data and downsampled_data_5m.
My ordinary queries look like this: SELECT /^usage_system$|^mean_usage_system$/ FROM "$retention_policy"."cpu" WHERE ("host" =~ /$server$/ and "cpu" = 'cpu-total')
So I have ability to switch value of variable $retention_policy and show data from different RP.
But sometimes there is a need to perform arithmetic operations with result of queries. For example, there is no metric in Telegraf that shows usage of inodes in percent: show field keys from disk name: disk fieldKey fieldType -------- --------- free integer inodes_free integer inodes_total integer inodes_used integer total integer used integer used_percent float
To calculate free inodes in percent I need to perform query like this: SELECT "inodes_used" / "inodes_total" * 100 FROM "disk"
For case described above I need to add regexp into query like this: SELECT (/^inodes_used$|mean_inodes_used$/ / /^inodes_total$|^mean_inodes_total$/ * 100) FROM "disk"
But regular expressions doesn’t work in queries with mathematic operations. Is it possible to add support of them?