Severity
MEDIUM-HIGH: CPU 99% + memory 95% + upward trend; service degradation likely, full outage risk if unchecked.
Root Cause
- Runaway process or memory leak consuming resources
- Sudden traffic spike or cascading request queue buildup
Actions
- SSH to affected host; run
top -b -n 1 | head -20 to identify top CPU/memory consumers.
- Kill or throttle the culprit process if identified; if unknown, restart the primary service gracefully.
- Check recent deployments, config changes, or traffic logs in last 30 minutes for trigger.
- Scale horizontally (add instances) or shed non-critical load immediately to relieve pressure.
- Confirm memory doesn't exceed 90% and CPU stabilizes below 85% within 5 minutes.
Watch
- CPU trend and absolute value (target <80%; alert if >90%).
- Memory usage (target <85%); watch for further creep toward OOM.
Escalate If
CPU remains >95% or memory >93% after 10 minutes of remediation attempts.