Anyone in IT will be familiar with the human tendency to only acknowledge what’s going wrong and be completely oblivious to what’s going right. When IT works well, everyone gets on with their job, taking the infrastructure, apps and services that they’re working on for granted. But when they fail? Well, then IT is very much under scrutiny.
Sys admins hate this, but are guilty of the same themselves. What gets their attention during their working day? The systems and services that are broken or being built. The rest is simply left to do what it should be doing; ticking over being the green dot on the network monitoring map.
But unwittingly, this can turn into an expensive behaviour. If the green dots don’t get attention, who knows what they’re doing? Whether they’re even being used anymore? Or whether they could be used for something else?
It’s possible, likely even, that these devices are whirring away in the data centre taking space and power and costly licences when they’re not serving any useful purpose at all. It’s time to become familiar with a new category of infrastructure: the ghost servers, VM creep and stranded storage capacity.
Finding the Needles in the Haystack
IT monitoring is usually thought of in association with underperforming services, but can as easily identify under-utilised services as well as over-utilised and failing ones.
We would recommend you approach the task by periodically reviewing a list of the systems you are monitoring, and use the following criteria to identify systems for closer inspection:
- Any servers that you or colleagues cannot put a function to
- Servers with consistently low CPU consumption – suggestive of little or NO activity
- Capacity on data disks which stays consistent over a long period of time (months)
- Non running VMs – identify owners and investigate need to retain
- Unmounted SAN storage LUNS
- Also watch out for devices that go down, but never appear to generate customer enquiries
