1. Run ps or top: Is there any process run away? (CPU usage > 50%)2. Check the load average with uptime, check if CPU is ever idle
with vmstat or sar -u3. Run vmstat (or sar -g): how much paging is going on?
( too many page-out: contention for memory)Q: Suppose you find a particular process is at fault, What to do?
A: Nothing!
Why? some processes are bound to slow down the system4. Restrict a process's consumption with limit
5. Run traceroute, ping, or netstat: is network overloaded?
6. Is server-related delays?