Scylladb process is more than 100% utilized even it is ideal

Hello,

We are facing high CPU utilization by Scylladb process even if it is ideal. And when we put a small load CPU jump to 400% and more.
Is there a way we can trace or enable advance logging to get this issue resolve.

Thanks
Sachin

Where do you see this high CPU utilization? In top or similar?

Note that checking ScyllaDB’s resource utilization in external tools is misleading. These external tools will report high CPU and memory utilization, even when the ScyllaDB is lightly loaded. To get a real picture of how much is ScyllaDB loaded, setup monitoring and check the load and memory metrics there.

Hi Botond,

Thanks for your reply. I am also surprise with the output from top command. Load average is 1.07, 1.65, 1.67 but in CPU% it says 146.5. Memory utilization is 47GB used as we configured it 48GB max. I will try to setup monitoring. If you can suggest anything will be appreciated.

top - 11:37:01 up 2 days, 23:23, 1 user, load average: 1.07, 1.65, 1.67
Tasks: 406 total, 1 running, 405 sleeping, 0 stopped, 0 zombie
%Cpu(s): 2.7 us, 1.1 sy, 0.0 ni, 96.2 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 64087.2 total, 10160.3 free, 48765.5 used, 5161.3 buff/cache
MiB Swap: 32735.0 total, 32735.0 free, 0.0 used. 13244.8 avail Mem

PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND

10251 scylla 20 0 16.0t 46.7g 109392 S 146.5 74.6 9721:10 scylla

We have our own monitoring stack (based on widely used open-source software), see ScyllaDB Monitoring Stack | ScyllaDB Docs for more information.

Thank you Botond, I will try ScyllaDB Monitoring Stack

It does seem unusual to have such a high CPU utilization reported by top while the load average remains relatively low.