5.Performance & Diagnostics

Diagnosing CPU saturation, run queues, context switching, hot code, affinity, and scheduler effects.

Interactive Linux process and host metrics for CPU, memory, load, task state, and quick triage.

Diagnosing resident memory, reclaim, page cache, swap, faults, NUMA, and application allocation behavior.

Sampling runnable/blocked tasks, memory, paging, block I/O, interrupts, context switches, and CPU states.

Diagnosing storage latency, throughput, queueing, cache behavior, filesystem, and application I/O patterns.

Diagnosing latency, throughput, packet loss, retransmission, queues, sockets, connection limits, and kernel networking.

Interpreting CPU and block-device throughput, latency, queueing, and utilization metrics.

An evidence-driven workflow across CPU, memory, disk, network, kernel, and application layers.

Visualizing aggregated stack samples to locate CPU or off-CPU hotspots.

Soft and hard per-process resource limits, inheritance, file descriptors, processes, and memory-related constraints.

Packet capture with BPF filters for debugging TCP, DNS, routing, retransmissions, and network failures.

A practical map of Linux commands for files, processes, networking, resources, and diagnostics.

Historical and sampled system activity for CPU, memory, I/O, network, and other kernel counters.

Per-process and per-task CPU, memory, I/O, fault, and context-switch statistics.

Tracing Linux system calls and signals to diagnose blocking, failures, and unexpected kernel interactions.