Research: How to display metrics effectively

scm-probe / scm-prom

Syscall prometheus exporter built using eBPF

0 stars 0 forks source link

Research: How to display metrics effectively #5

Open utkarsh-1905 opened 2 weeks ago

utkarsh-1905 commented 2 weeks ago

Points to consider (as of now):

It should show which system calls are most used
Rate of system calls per time interval

Should we use Grafana to graph this out or some other tool, considering we have to write queries using PromQL

utkarsh-1905 commented 2 weeks ago

I haven't figured out a way to find the most used calls out of all syscalls; considering there are upwards of 300 syscalls on the latest Linux kernel, should we change how we send metrics to the server?

utkarsh-1905 commented 2 weeks ago

If not Grafana, can we make our own page to collect metrics from Prometheus and create custom dashboards alerting mechanisms