TechTip: Amazon S3 Outage…Your data is unavailable…Nope not their fault
Naysayers beware! Pointing fingers may just bring the spotlight back to you and your lack of a continuity minded architecture.
Yes, yes, yes. Amazon experienced a service outage yesterday. It IS a big deal…for those customers that did not heed the warnings and listen to the instructions from Amazon themselves. While yesterday’s outage was a big deal for those affected, it’s important to remember that Amazon S3 has proven remarkably dependable over the years, and while the Northern Virginia datacenter (us-east-1) was down, S3 remained up in its 13 other regions.
Technology WILL fail…hard drives, mainboards, switches, routers…everything has the potential to fail. Which is precisely why, when we architect a solution, it is created with as much resiliency as possible. Prepare for the inevitable. This is just as true of “Cloud” providers as it is for your local infrastructure. This is EXACTLY why the cloud makes sense. These features are built into every service offered by AWS and Azure. Cloud providers are not infallible, and as I said, they profess as much when you are creating your environment. “Build across different availability zones.” “Extend your content across regions.” You wouldn’t just use one instance or one datacenter location if you were housing your applications yourself, so why would anyone think that the same logic should not be applied here?
This is not an I told you so moment, but hopefully a learning experience regarding the fact that we need to architect our cloud environments just as soundly as we would if we were racking and stacking it ourselves. Public and Private cloud both offer unique flexibility and in some cases, cost saving measures, but both need to be built mindfully and responsibly.