goebpf attaches eBPF programs to Linux scheduler tracepoints
(sched_switch, sched_wakeup, block I/O, syscalls, and network events)
and turns raw kernel activity into a ranked table and a plain-English
status line, live.
What it does
monitor– live context-switch tracing with a STATUS line that’s always emitted, even when the system is healthyexplain– plain-English interpretation of sleep stacks: what a process is waiting for, and whether that’s normaldiagnose– a one-shot 30 second health check combining monitor and explain, with suggested mitigationscompare– diff two saved JSON reports to see what changed after a config change--serve– a background HTTP server exposing Prometheus/metricsand a live/reportendpoint in json, yaml, or text
Quick start
go generate ./...
./compile
sudo setcap "cap_bpf,cap_perfmon,cap_sys_resource=ep" ./goebpf
sudo ./goebpf monitor