Field notes for diagnosing Linux applications, containers, and performance issues with system call tracing.
strace reveals exactly which futex or recvfrom call your process is blocked on — something no APM or log aggregator can show.
Identify tight loops of failing stat() calls or excessive polling that monitoring dashboards only show as "high CPU".
Quickly diagnose ENOENT, EACCES, missing libraries, and port conflicts in Kubernetes pods where logs are empty.
What is strace, ptrace internals, syscall lifecycle, blocking vs non-blocking.
Production incidents where logs/metrics fail. Comparison tables and real scenarios.
Attaching to processes, tracing children & threads, timestamps, durations, output files.