Ask your customers.
If it’s a customer-facing application, it doesn’t matter if the system is down or just slow. Customers will abandon you and go somewhere else.
Configuration mistakes like the DNS issue Apple experienced took iTunes offline to the tune of $25 million in potential revenue losses in 12 hours, according to one media report. Customers were upset. Apple was unable to capitalize on expected demand. The impact was felt globally. Activity ceased for several hours. That’s an outage of magnificent proportion.
Or take the grandaddy of slowdowns - healthcare.gov. The issues were largely traced to a case of bad planning - capacity planning. When the site went live, millions of uninsured Americans rushed to the website only to find they had to wait, as some media outlets reported, in virtual ‘waiting rooms’ of 40,000. Access unavailable? Reputation damaged.
You have to provide more insight into workloads, performance and resource consumption in large scale environments that include hundreds or thousands of servers. Provide more in-depth and predictive analysis to help your organization better predict and prevent glitches and mishaps.
What does utilization really mean to your customers? To your business? Use latency as a better performance indicator. Analyze the components of response time (i.e., latency) so you can see how much time a specific workload transaction, application, virtual machine or system is actually getting service versus waiting for a resource.
Anticipate the impact of new application rollouts. Perhaps it isn’t practical to test with production-level systems and loads, but it’s important that the planned architecture is analyzed using methods that can accurately predict how it will perform.
Requirements change. If you’ve tested performance and planned properly and then have a last-minute requirement change, test again. You shouldn’t skip testing in order to meet a target launch date. Proper testing can reveal your ‘real’ capacity requirements before production and potentially save you time, money and public embarrassment. Like Forrester’s JP Garbani once said, it’s cheaper to prevent a problem than it is to fix it.