Instance became intermittently unresponsive. SSH access was timing out, EC2 Serial Console failed to connect, and stop operation was unusually slow. Status checks showed insufficient data during the event.
Initial investigation after regaining SSH showed no memory pressure, no swap usage, and no disk errors. CPU credit balance remained healthy at over 500 credits. However, vmstat revealed high CPU steal time, peaking above 50 percent, indicating significant hypervisor level contention.
No e...