htop Cheat Sheet: System Monitoring & Troubleshooting

A concise reference for interpreting htop metrics, color codes, and shortcuts.

1. Essential Shortcuts

Key Action Context
F3 / F4 Search / Filter Locate specific processes (e.g., python, node).
F5 Tree View Visualize parent-child process hierarchy.
F6 Sort By Reorder by CPU%, MEM%, or PID.
F7 / F8 Nice - / + Shift priority (Lower NI = Higher priority).
F9 Kill Send signals (e.g., SIGTERM, SIGKILL).
F10 Quit Exit htop.

Note: On macOS or laptops, use Fn + Fx or Esc + [Number] (e.g., Esc + 3 for F3).

2. Meter Color Codes

CPU Usage

  • Green (User): Standard user-space processes.
  • Red (Kernel): Kernel-level tasks (I/O, system calls, interrupts).
  • Blue (Nice): Low-priority tasks (will yield to standard processes).

Memory (RAM)

  • Green: Used memory (actually allocated by apps).
  • Blue (Buffers): Temporary storage for raw disk blocks.
  • Orange (Cache): Cached files to speed up disk reads.

Note: High cache/buffer usage is expected behavior in Linux; it is released automatically when apps demand RAM.

3. Load & Tasks

Tasks: <tasks> / <threads> / <running>

  • Tasks: Total process count.
  • Threads: Total execution threads.
  • Running: Units currently using or waiting for a CPU core.

Load Average (1 / 5 / 15 min)

  • How to read: Compare values against your logical CPU core count.
  • Healthy: Load < Core count.
  • Saturated: Load ≥ Core count (processes are queuing for CPU time).

4. Process Columns

Column Full Name Description
PID Process ID Unique system identifier.
NI Nice Value Priority offset (-20 to 19).
VIRT Virtual Mem Total mapped address space (includes shared libs).
RES Resident Mem Physical RAM used. The best indicator of actual memory footprint.
S State R (Running), S (Sleeping), D (Uninterruptible Sleep/Waiting for I/O).

5. Quick Diagnostic Scenarios

  • High Load + Low CPU%: Usually indicates an I/O Bottleneck. Check for processes in the D state.
  • Sustained Red CPU Bars: Suggests excessive kernel overhead, driver issues, or high system call frequency.
  • Swap Usage Increasing: Physical RAM is exhausted. System performance will degrade due to disk latency.
  • Shifting PIDs: If a process PID changes constantly, the service is likely crashing and restarting in a loop.

Config: ~/.config/htop/htoprc

Developing