Many column inches were written about the recent Amazon AWS outage in the US when emotions were high and tempers were frayed. Now that a few weeks have passed, we thought it was timely to look at it again and discuss what lessons have been learned by cloud computing customers and by businesses like Digital Mines.
Amazon provided a very detailed explanation of what went wrong and why. A combination of user error, process deficiencies and implementation errors meant that a simple mistake caused a catastrophic snowball effect on storage volumes know as EBS. This problem took out many web sites and services based in their US-East region. Those who were relying on the independence of Availability Zones in a region to protect them were dismayed to discover that it didn’t work as planned.
The outage highlights a critical lesson about any form of web hosting – it takes expertise to deploy and configure a resilient robust setup. Unfortunately many businesses have made the mistake of thinking that once they have moved to the cloud, all their problems will be automatically taken care of. That is not the case at all.
Amazon provides an incredibly rich and deep set of services in AWS. But they leave it up to you to configure it correctly. EBS volumes are generally extremely reliable, EC2 instances generally behave well, S3 is rock-solid for object storage in most cases. But is that enough for your business? If your EBS storage volume that holds your database is accidentally deleted, do you have snapshots of it? Do you have database backups on S3 for dealing with catastrophic failure? And if an entire Amazon region goes down, how are you going to get back up and running? Do you just wait for Amazon to sort it out?
These are the types of questions anyone with a web-site on any type of hosting should be asking. If you are using the cloud then you need to be sure that you have a robust system and you need to know what the escalation path is when problems occur.
That’s why fundamentally, you need companies like Digital Mines. We provide the Amazon infrastructure, we make it easy to manage and maintain, but most importantly of all, we have the expertise to put in place a configuration that matches your budget and your needs.
Simply ask yourself this “what would I do if my Amazon server disappeared tomorrow?”. If you can’t answer that, then you need to be talking to a cloud services provider.