Industry
eCommerce
Annual Revenue
$1.4B
Use Cases
Load Testing
Regression Testing
API Observability
Overview
Get started today
Replay past traffic, gain confidence in optimizations, and elevate performance.
National Restaurant Chain Gains Confidence in App Performance During Peak Hour Traffic
- Industry: eCommerce
- Annual Revenue: $1.4B
- Use Cases: Load Testing, Regression Testing, API Observability
Company
As one of the largest quick service restaurant chains in the United States, this fast food enterprise holds its dedication to quality food and top customer service in high regard. Investment in powerful technologies plays a big role in their ability to continually delight customers. As the Senior Principal Team Leader for Platform Engineering explains, “Like many other consumer and ecommerce companies, we are constantly updating and evolving our digital platform so we can better serve our customers and improve daily operations. The only challenge with this is, it’s on us to make sure those services and applications are performing perfectly—even during peak meal times when traffic can surge to 300,000 requests per minute!”
Ensuring the performance and stability of their digital platform is no easy feat, especially across many engineering teams, each pushing code changes and new features and functions daily to support the business. This pace of innovation, coupled with the complexity of the distributed Kubernetes environment that relies on many integrations and third party services, makes it difficult for teams to gain visibility into all dependencies and understand how one change might impact an application or service.
The missing piece: automated, scalable testing in CI/CD
“Our platform engineering team thinks of themselves like a ‘restaurant in the cloud’. We support the engineering teams who oversee digital services for customers, staff and store operations. These developers are moving fast every day to improve experiences and yet, there hasn’t been a consistent, reliable and scalable way to test code before production. This essentially means we don’t have 100% assurance that the changes or new services are going to perform,” the Principal Team Leader explains.
Until recently, developers were using a mix of methods to test changes and mock dependencies. Some teams were testing in production, others were attempting to test pre-prod but were struggling to accurately replicate and map all the dependencies. Some were using mock libraries with only limited success. All this meant that problems would often present themselves in production.
The Principal Team Leader recalls one particular incident when a third-party provider went down during the lunch hour rush, meaning that the company could not process orders for a period of time. “When something like this happens, it has a big impact on the business. It was all hands on deck to find and fix the issue, but this was one of the catalysts for finding a new, more consistent way of testing.”
Pressure surged when digital sales jumped 50%
Additional pressure hit the platform engineering teams when online sales jumped 50% during the COVID-19 pandemic. As the Principal Team Leader elaborates, “When we saw the sales surge overnight, we knew we had to focus on the resiliency of our digital platform. Like most other companies our size, we’d accrued some technical debt and the new pressure on our system was exposing weaknesses. Our teams continued to pour new features and functionality into the platform—which was great—but with the new levels in traffic, we knew we needed to find a consistent, scalable way to test and mock dependencies with real world scenarios and data across all engineering teams. That’s when we reached out to Speedscale.”
“Our teams continued to pour new features and functionality into the platform—which was great—but with the new levels in traffic, we knew we needed to find a consistent, scalable way to test and mock dependencies with real world scenarios and data across all engineering teams. That’s when we reached out to Speedscale.”
One solution checked all the boxes
The quick service restaurant’s IT environment is a distributed microservices architecture, with millions of monthly users. Complexity is widespread, so when it came to finding a new solution, the platform engineering team wanted to find something that could handle the intricacies of their environment and scale across the business. They also wanted something that was easy to maintain and become part of the developer workflow.
Talk to an expert today
“In addition to Speedscale, we looked at a few other paid and open source solutions, but they couldn’t do all the things we wanted, they lacked reporting and made it very labor-intensive to dynamically change timestamps or authenticate tokens. Speedscale was customizable to our traffic, and our developers use it to stress test our code changes with the click of a button. The Speedscale team has also been exceptional to work with. We’ve built trust quickly because they could support our processes and platform engineering philosophies.”
“The Speedscale team has also been exceptional to work with. We’ve built trust quickly because they could support our processes and platform engineering philosophies.”
Increasing confidence in app performance
Speedscale is now the preferred way the platform engineering teams test—by running traffic replays, and mocking dependencies automatically as part of their CI/CD process.
“We’re basically standardizing testing on Speedscale. It’s recording everything going in and out of our platform and automatically generating simulated production transactions and mocking scenarios based on real traffic and up-to-the-minute usage scenarios,” explains the Principal Team Leader. “It’s part of our teams’ workflows. We know how changes and updates are going to impact the performance of our applications and services before they go live. We have more confidence from being able to test accurately and at scale.”
“It’s part of our teams’ workflows. We know how changes and updates are going to impact the performance of our applications and services before they go live. We have more confidence from being able to test accurately and at scale. The value is undeniable.”
As a result of implementing Speedscale, developer productivity and efficiency has also improved significantly, supporting the company’s core value around employee work/life balance. “Looking after employees is central to our philosophy. Because we have full visibility of our environment and confidence when we push code live, we’re reducing the amount of stress and urgent work placed on our developer and engineering teams.”
The team has since expanded its use of Speedscale to prepare for expected traffic spikes from upcoming promotional campaigns.