The data deletion was done inadvertently by a maintenance process run accidentally by one of the few developers with access to this area and the mistake was not noted at first. When the problem started, Netflix started focusing on the API errors, but it took some deep digging to find the root of the problem.
It wasn't until 3:05pm EST the next day, which was Christmas, that Netflix reported that the service was up and running. To make sure something like this doesn't happen again in the future, changes have been made to prevent accidental modification without approval. Additionally, Netflix says it has learned how to get the service up significantly faster in the unlikely event that the same events happen again.
source: AmazonWebServices via Electronista
No comments:
Post a Comment