AWS Outage Takes Online Stores Down on Valentine’s Day

14 Feb 2019 3 min read

In the early hours of Valentine’s Day many shoppers in search of last-minute gifts were instead met with sites that were down, due to an unexpected and rare incident with Amazon Web Services that impacted thousands of online stores.

Amazon Web Services (AWS), a cloud services platform offering computer power, database storage and content delivery, began experiencing database problems in the Asia Pacific region at around 6.00am (AEST) this morning, reporting “increased create times and cluster unavailability” on their status page.

This caused widespread outages for many sites across Australia, including Big W, Afterpay, leading tech platforms MYOB and Zendesk, as well as a number of Neto sites. Amazon resolved the issues at 8.49am and sites began recovery soon after.

Neto’s VP of Engineering, Justin Hennessy, says that Neto is built to minimise the risk of these type of events and AWS incidents like this are extremely rare.

“The Neto platform is designed in such a way that we have high resilience and redundancy with infrastructure for Neto stores spread across multiple availability zones.”

“This is the first incident for this service that we’ve seen and we worked closely with Amazon through the resolution period.”

Neto and AWS

The fifth largest software company in the world, AWS is the leading provider of cloud infrastructure and has 34% share of the market, ahead of competitors Microsoft and Google.

Neto moved to AWS two years ago after identifying that the previous infrastructure we were using was unscalable and restrained by manual processes. Instead of buying rack space in a data centre and setting up our own physical servers, AWS allows us to fully automate processes that help us configure and scale virtual servers to our needs.

For our customers, using AWS means more reliable performance and security, without interruptions to their service and enables us to roll out new features continuously.

One of the AWS services that Neto uses is a Relational Database Service (RDS) called Aurora, which allows us to store customer data, and this was the service that was affected in today’s incident.

“All cloud providers are built to be highly resilient, but there are occasions that there are interruptions to service, that are very rare”, Justin says.

Insuring your business against the unexpected

While we take site stability very seriously here at Neto and try to make the platform as resilient as possible, unexpected incidents and outages like today’s can occur. So what can you do to protect yourself?

Many physical businesses have insurance to protect their property against unexpected disasters like floods, storms, or break-ins. Many also have another type of insurance called business interruption or continuity insurance, which protects the revenue stream when the business is forced to close during this period.

National Product Manager for Commercial Lines, Andrew Norris in an article for QBE insurance, explains:

“Business interruption insurance deals with the continuity of income after insured property is damaged or lost after a catastrophic event.”

“It operates to cover business cash flow and in the process also protects the owner’s income stream and their investment in the business. The cover pays for ongoing costs that continue regardless of whether the business is closed or substantially affected; e.g. rent, fixed electricity, gas and water charges. It also covers additional costs incurred to minimise the effects of the closure or downturn, plus net profit and/or loss."

And just as bricks and mortar businesses can protect themselves against these disasters, online businesses should be protecting themselves and their income against the online equivalent.