Performance & Resource Troubleshooting

What is it?

Performance troubleshooting finds which resource is the bottleneck — CPU, memory, disk I/O or network — using tools like top, free, iostat and vmstat, then ties it to the responsible process.

Why it matters

'The server is slow' has many causes; guessing wastes time and can make it worse. Isolating the actual bottleneck is what turns a vague complaint into a fixable problem.

Where you see it

`top` (CPU/mem per process), `free -h` (memory), `iostat` (disk), `vmstat` (swap/IO) — read together to localize the constraint.

What normal looks like

Load proportional to work, memory used but not swapping heavily, disk I/O not saturated — headroom on the constrained resource.

What suspicious looks like

One process pinning 100% CPU, constant swapping (memory exhausted), or iostat showing the disk at 100% utilization — each a distinct, fixable bottleneck.

How analysts investigate

Localize before acting: which resource is saturated (CPU/mem/disk/net), then which process owns it, then why — never change tuning before you know the bottleneck.

Common beginner mistakes

Adding CPU when the real bottleneck is memory or disk I/O.
Reading one tool (top) alone and missing an I/O or swap problem.

Find the bottleneck first

  'server is slow' → localize the resource:
    top    → java pinning 100% CPU?  (CPU-bound)
    free   → 0 free, swap climbing?   (memory-bound)
    iostat → %util 99 on sdb?         (disk-bound)
  attribute to the process, then find WHY — don't guess the fix

Isolate which resource is saturated and which process owns it before changing anything.

Quick check

A box is slow; free shows swap climbing and 0 free memory. The bottleneck is...

A quick self-check — it doesn't affect your XP or progress.