And it is that we we know that did you hear about the problem which affected the network last February 28, in which after a little more than five hours the service and some applications were affected by this fall. In this case the problem has the base service S3 Amazon Web Services (AWS) and full affect a handful of services like IFTT, GIF Giphy, Trello or Hootsuite web among many others that are hosted in this Simple Storage Service from Amazon.
For now we have clear once already solved the problem, is that after research conducted by Amazon, the cause of the problem has been a human error. Some workers of Amazon S3, performed maintenance in the billing system and required the closing of some servers, as you can be thinking this all ended quite badly and is wrongly closed more servers needed for the task and the necessary subsystems were not able to recover information, by which the service stopped working.
In this sense and to see the magnitude of the problem the only thing they could do was to just restart the entire system and obviously not done in minutes since many processes require more time due to the amount of information stored. In this regard in addition to this, many servers had never previously been rebooted and this further affected the reactivation of the service.
Now we doubt about if this can be repeated again since while it is true that the engineer in charge of the task of maintenance in the billing system was right about the manual, but mistakenly touched something that should not cause the fall of the network, could happen again in the future. On the other hand, and after the problem a new safety option now exists in which engineers may not disable servers and the dashboard will be a separate system to the S3 to avoid the same happening in a hypothetical future maintenance task.