Hi.
Here is what I have on amazon ec2 instance:
This is stacked graph of cpu metrics (which are system.cpu.util[,user,avg1], system.cpu.util[,system,avg1] as so on) with max y = 100.
As you can see, all metrics except interrupt sometimes (often on :00 minutes, but not always) jumps to 100. I've catched the moment, top during that jump reported normal values (0-5%).
What might be the cause?
I've already checked - there is no ntpdate on cron scripts on the machine, ntpd is running and reporting no problems.
Here is what I have on amazon ec2 instance:
This is stacked graph of cpu metrics (which are system.cpu.util[,user,avg1], system.cpu.util[,system,avg1] as so on) with max y = 100.
As you can see, all metrics except interrupt sometimes (often on :00 minutes, but not always) jumps to 100. I've catched the moment, top during that jump reported normal values (0-5%).
What might be the cause?
I've already checked - there is no ntpdate on cron scripts on the machine, ntpd is running and reporting no problems.
Attached Files