Technical Blog


Technical and Product News and Insights from Rackspace

Scaling AWS for Black Friday: Best practices


Originally published on October 21, 2019 on Onica.com/blog

With the holiday season approaching, your online shopping site is likely to experience high volume traffic because events such as Black Friday and Cyber Monday sales are on schedule for their annual return. Black Friday raked up $6.22 billion in online sales in 2018, and with a sustained annual growth rate of between 20-25%, the 2020 holiday season is shaping up to be one of the biggest online shopping events yet.

While this may be an exciting time for shoppers, events like Black Friday and Cyber Monday can be challenging for eCommerce sites and online businesses, who need to ensure that their online infrastructures can scale and maintain strength under abnormally high traffic loads. Following problematic instances from earlier years, 2018 saw many large retailers face website crashes, including high-profile retail brands, with some having lost over $700,000 in sales per estimates by experts.

On the entertainment front, Avengers: Endgame broke box office records in ticket sales, crashing the websites of AMC Entertainment®, the largest U.S. theatre chain, as well as Fandango®, a popular ticket sales website. Despite the recurrence of such crashes, substantial growth in online sales makes it hard for companies and eCommerce platforms to accurately project and adequately prepare for these high traffic events. Furthermore, customer frustration that grows out of these crashes can have a drastic impact on long term sentiment, resulting in reduced sales and a potentially tarnished company image.

In light of these implications, it is important to start thinking about addressing your website or applications’ scalability and stability. In fact, scalability is one of the primary reasons businesses migrate their infrastructure to the cloud. AWS® offers a highly supportive environment for such high traffic events, with a variety of tools and features that can ease efforts in this regard.

Rackspace Onica has helped a broad set of customers move their infrastructures onto AWS for exactly these considerations. MovieTickets.com, a ticket supplier, moved their entire infrastructure on AWS to support the high scalability that they expected to need during the release of Star Wars: The Force Awakens. This move allowed them to avoid crashes similar to the ones that Fandango and AMC Entertainment faced when they released Avengers: Endgame. The Orange County Registrar of Voters also sought Onica’s help to migrate their website to AWS so that they could support high traffic during election times.

Five AWS auto-scaling best practices to build resiliency against high-volume traffic

Migrating to the cloud, although a great first step, is not the only measure required to ensure that a company’s website or application can survive unprecedented traffic surges. In addition to automating and scaling as much as possible, the following set of AWS auto scaling best-practices can help companies maximize the robustness and preparedness of their infrastructures.

  1. Place checks and measures with Amazon CloudWatch. Monitoring the performance of critical assets during high traffic events such as Black Friday sales is very important. A workflow where the right people are alerted or appropriate actions triggered as soon as possible, through alarms and notifications set up for key performance metrics, can help you stay on top of issues if they arise.

  2. Use AWS Auto Scaling. It is good practice to place the most needed resources into auto-scaling groups. These groups can help you treat multiple instances as a single object and scale specific instance types with specific resources together. AWS Auto Scaling also performs regular health checks on Amazon EC2 instances, replacing those that fail the check. In addition to supporting PCI-compliant processing, storage, and transmission of credit card data, these features make AWS Auto Scaling an effective tool to combat high-traffic load.

  3. Leverage Elastic Load Balancing. Distributing incoming application traffic across multiple instances in your AWS Auto Scaling group can ease traffic load. Elastic Load Balancing can help you achieve this automatically, in addition to balancing traffic across multiple Availability Zones.

  4. Amazon Route 53® is typically used to route DNS queries to your load balancer, and you can also configure it for DNS failover. Amazon Route 53 then checks the health of registered Elastic Load Balancing endpoints to determine your application’s availability and routes requests to the most available resource.

  5. Using Amazon ElastiCache can help improve performance by storing information on fast, managed, in-memory stores to accelerate data retrieval compared to slower disk-based databases. Additionally, it detects and replaces failed nodes and reduces the risk of overloaded databases to keep website and application load times speedy.

Ensuring all the preceding measures are in place is only one of the early stages in preparing for high-traffic events. Forecasting and building future-traffic models based on historical metrics to estimate your future resource requirements is a very important consideration in being prepared. This data can inform test scripts that test the robustness and scalability of your infrastructure. Coupling these tests with surprise drills for the team can help build preparedness and reveal potential areas of concern. AWS also allows vulnerability and penetration testing as long as it falls within the AWS Security Testing Terms and Conditions.

In the end, the goal of scaling for such high traffic events is to ensure that your customers have a great experience as they interact with your application or website. A slow or unavailable network can cost more than an immediate sale, as customer sentiment is affected in the long term. Hence, thinking about your customers' needs and wants while fortifying your infrastructure can go a long way in ensuring you spend your effort in the best places.

While these tools and best practices serve the scalability and stability needs of high traffic periods, you should be prepared all year long. Being aware of current environmental and economic factors, shopping traffic trends, and buyer behavior throughout the year can help build a clearer picture of what you can expect when the shopping season commences.

Learn more about our AWS services.

Use the Feedback tab to make any comments or ask questions. You can also click Sales Chat to chat now and start the conversation.

post avatar
Matt Charoenrath

Share this information: