This is (the start of) a collection of useful eBPF scripts for instrumenting HPCs
You will need to install bpftrace and the appropriate Kernel headers on your system. On Red Hat-likes:
dnf install bpftrace kernel-headers-$(uname -r)
These scripts are intended to produce output that can be easily scraped by shell tools and forwarded to central collector systems (such as Telegraf).
You can either run the scripts as standalone scripts (i.e., ./nfsd.bt) or invoke the bpftrace tool (e.g. bpftrace nfsd.bt).
These sample the various tracepoints for a short interval, such that they can be executed on a cron or another similar tool.