virtualization memory management
Posted on 2014-11-14
How do you counter your application team or 3rd parties when they require systems to be set up according to ridiculous hardware or load specific requirements such as 24 or 32GB of RAM, although the application runs fine on the provided 8GB. But still they point their fingers because the system is not tuned to what the software developers state as minimum or recommended settings.(Usually this is based on some load in a large enterprise environment, which may not be the case).
I do refer to the systems utilization, weekly or monthly graphs of memory usage, but always when they have a problem it's because the system does not have enough memory allocated(But 20-30% of the VM's memory is not allocated by the server, even when configured below what the application developer states as minimum or recommended.)
I am a big fan of right sizing VM's to lower the overhead, to optimize and to get the most out of the environment.
Therefore, what are your thoughts on this and how do you handle such cases. I know from a business perspective(you migh't say), either the customer pays for the hardware or pays for the resources used, but from an operational perspective it is still my task to keep resources well spent and right sizing the virtual machines in "my" environments.
I know at the end of the day it's the same people who ask me, when the environment is capped on resources and they have to spend money on more. What can we consolidate or save resources on?