Clearly, uptime is the most important operational metric for a data center. The most common reasons for downtime are: mechanical A/C failures, power loss and network connectivity problems.
Data center outages tend to be public affairs, and stories about downtime are easy to find. Some better known data center power failures include:
- Verizon DC power failure in January 2016 caused 9 hours of downtime, affecting flights throughout the USA.
- GitHub DC power failure January 2016 caused 2 hours of downtime
- Fujitsu DC power failure August 2015 causes problems ‘for days’
A firm that tracks power outages, Eaton, estimates that outages cost the US economy $150 billion in annual damages.