I have a worker tier that is using PHP process control (pcntl) functions to send multiple requests to the SNS service in parallel. My question is how best to tune MPM_PREFORK directives to get the best out of the box and how to calculate the most concurrent PHP processes I can run with the resources available. I am running a T2 medium at the moment which is spec'd like this:
Model |vCPU | CPU Credits / hour | Mem (GiB) | Storage
t2.medium | 2 | 24 | 4 | EBS-Only
Also if anyone could shed some light on how the CPU credits are used/factored into this calculation I would be very grateful. I have currently sort of applied the thinking in these posts..
http://cloudinservice.com/tune-apache-performance-using-mpm-prefork-module/
http://fuscata.com/kb/set-maxclients-apache-prefork
Any help much appreciated