scheduling group cpu metrics are summed across shards

[avikivity]

This shows that e.g. statement is consuming 8 seconds of CPU per second. You can only understand it if you know how many shards you have.

[amnonh]

Can you please include the panel, dashboard,Scylla version, and Scylla-Monitoring version you are looking at?

[amnonh]

@avikivity can you please add the info?

[avikivity]

It was 4.0, the CPU metrics.

[amnonh]

@avikivity what scylla monitoring version?

[amnonh]

I'll add an optional aggregation function like some other dashboard have and set the default to avg

[avikivity]

@amnonh I don't know, I didn't start the system.