Hi
Long story short, we have to run an ec2-instance to serve 5K connections per second. We tried to take bigger and bigger instances but for some reason the vmstat always showing the machine peaking the kernel-time with the CPU cores being at 100% usage.
We are testing with a php script that just does <?php echo "test"';?>
we did increase the php-fpm to 5000 (workers) and the same for apache (we are running apache in event-mode).
Any suggestions on what to tune in the centos-machine itself?
Thx.
Long story short, we have to run an ec2-instance to serve 5K connections per second. We tried to take bigger and bigger instances but for some reason the vmstat always showing the machine peaking the kernel-time with the CPU cores being at 100% usage.
We are testing with a php script that just does <?php echo "test"';?>
we did increase the php-fpm to 5000 (workers) and the same for apache (we are running apache in event-mode).
Any suggestions on what to tune in the centos-machine itself?
Thx.
Last edited by a moderator:
