Episode 51 of Linux Server Admin with Fexingo dives into BPF (Berkeley Packet Filter) as a practical tool for real-time performance tracing on Linux servers. Hosts Lucas and Luna walk through a concrete scenario: diagnosing a mysterious latency spike in a web application without restarting or installing new packages. Lucas explains how bpftrace lets you attach one-liners to kernel probes to measure disk I/O, syscall latency, and scheduling delays, using real examples like tracing ext4 file operations and identifying a slow NFS server. The episode covers the difference between classic BPF and eBPF, common one-liners for sysadmins, and safety mechanisms that prevent crashes. Listeners learn how BPF can replace older tools like strace for production use by being safer and more efficient. The hosts also touch on how to check if your kernel supports BPF and where to find pre-built tools like BCC. Perfect for server admins who want deeper observability without overhead.

#Linux #BPF #eBPF #PerformanceTracing #Sysadmin #Bpftrace #Kernel #Observability #ServerManagement #LinuxPerformance #BCC #Troubleshooting #Technology #FexingoBusiness #BusinessPodcast #ServerEngineering #DevOps #RealTimeMonitoring

Keep every episode free: buymeacoffee.com/fexingo