Photo: mybulldog via Flickr
Netflix runs one of the most complex cloud computing architectures to stream all those movies to all those devices.How does it make sure it works all the time?
With an army of monkeys.
They’re not literal monkeys. They’re computer programs that test stuff on the Netflix architecture all the time to make sure it’s working properly, the company explains on its blog.
The first monkey was called Chaos Monkey and it worked by randomly shutting down a part of the Netflix architecture to see if it would still hold up. The idea was to have the equivalent of a real monkey loose in their data centre, breaking stuff and chewing through cables.
Since Chaos Monkey worked well, they now have Latency Monkey, Security Monkey and plenty of other monkeys, including Chaos Gorilla, which is an even more devastating version of Chaos Monkey.
It’s fun, interesting from a technical perspective, and shows how Netflix’s vaunted management works in practice.