Diagnostic model

Start with symptoms, establish current system state, inspect the relevant subsystem and validate the smallest change that resolves the issue.

Observe → Reproduce → Inspect → Change → Verify

Useful first checks

systemctl --failed
journalctl -p warning..alert -b
ss -tulpn
free -h
df -h

Commands are starting points, not conclusions. Correlate their output with timestamps, recent changes and application logs before deciding on remediation.