Quantifying Hard Interrupt Handling Latency and Frequency with Hardirqs

Hardirqs measures the duration of hardware interrupt service routines by attaching BPF tracepoints or kprobes to irq_handler_entry and irq_handler_exit kernel hooks. It records entry timestamps per IRQ vector, calculates total handling time upon exit, and aggregates latency distributions into logarithmic histograms directly in kernel BPF maps.

Infrastructure operations and low-latency networking teams use hardirqs to isolate hardware-driven performance bottlenecks. When servers experience unexplained packet drops, audio or video streaming jitter, or storage frame losses, hardirqs determines whether specific network interface cards or NVMe controllers are consuming excessive CPU time inside interrupt handlers.

The overhead of hardirqs is proportional to the overall interrupt rate of the hardware devices on the machine. Under extreme network packet storms or high disk IOPS, attached kprobes can slightly increase CPU overhead, though in-kernel BPF map aggregation minimizes impact compared to user-space logging. Access requires full administrative privileges due to hardware-level instrumentation.

> System.Execute()

sudo hardirqs-bpfcc -d -T 1 10

> Architecture & Security Context

Verified for Valters IT Hub. When deploying this configuration in a production homelab or enterprise environment, ensure that you monitor system resources and implement strict access controls. Administrator privileges are required.

ID: VALTERSIT-EF12093DBC48