AWS Outage: Things we can all learn

After large swathes of the internet were plunged into chaos by an Amazon Web Services (AWS) outage last night, we look at the things we can all learn from the technical difficulties experienced by companies of all sizes.

Posted on 01 March 2017 -
Tibus BY Tibus

The cost

First and foremost, anybody affected by the AWS outage - either because it took down your own website or a service on which your business is reliant - needs to put a price on the cost of the outage. Giving an actual figure to the disruption might provide context for the relative expense (or lack thereof) of putting in place a new or improved contingency plan.

The fact that AWS hosts a lot of stuff

Yes, we all knew that AWS was the big fish in the web hosting pond, but last night’s disruption really brought home how reliant we all are on Amazon’s infrastructure. The new level of awareness of that reliance can only be a healthy thing for forward planning.

The rookie error

We’re not going to get on Amazon’s case over the outage. It happens to all web hosts from time to time and anybody who claims to be above all that is telling fibs. And it generally happens when changes are being made that will improve services in the longer term. AWS generally does a very good job at an almost unimaginable scale, which makes their rookie error all the more surprising: their system status page was hosted internally on the same infrastructure that the outage affected, which meant the status page was also unavailable. Well worth bearing in mind if you need to update customers during similar tech problems.

The redundancy and failover option

AWS will fail again (we’ve already established that all hosting infrastructures will). If that caused you major problems - or if you discovered that AWS’s in-built availability zones didn't provide the level of backup you were expecting - then it is worth considering a redundancy and failover plan with a completely separate provider as an additional safety net. Of course, we’d be happy to help if you’d like to explore some of those options with us.

Equally, if you’re with a host other than AWS and your provider hasn’t been on your case about a redundancy and failover option, it might be time to move on. You might not need one, but your web host should at least be discussing this with you.

Interested in that extra safety net?

If you want a new redundancy and failover contingency in place before AWS fails again, talk to us.

Get in touch