High cpu utilization


I am seeing very high CPU utilization when grafana reads from graphite data source.I have large server running 32 cores of CPU and 60Gig of memory with plenty of disk space.

The issue I am seeing is that during reads from grafana to the graphite’s default time-series database, the CPU fully consumed to a point that Grafana queries times out.


I am using Grafana v4.5.2 (commit: ec2b0fe).

Any help would be most appreciated.


In the above screenshot, I am also not sure why Grafana starts 3 apagche2 processes to read the data from the datasource.

The retention schema that I’ve set up stores metrics every 1s for 7 days and 30s thereafter from 7 days to 3 months. Could reading these many data points when I run a report for 3 days or more cause of the high CPU?