First, Kong’s servers were not shut down. They were running the entire time. It was a DNS problem, which prevented people from finding the IP address to the servers.
Second, the DNS provider didn’t experience an outage, they erroneously deleted Kongregate’s entries (and others, too lazy to look up the reference). Redundancy couldn’t really help there because to most computers, this doesn’t look like a “DNS outage”.
Third, even after fixing it the nature of DNS propagation means that not every part of the world will observe the fix at the same time. This is because DNS isn’t maintained in a single central location that every computer on the planet connects to; rather, it is distributed throughout DNS servers around the world, and these DNS servers replicate changes to each other.