Adopting DevOps methodology is more than hype. In today’s ever-changing business environments, the ability to react fast and scale up is essential.

Thirty-two percent of customers will walk away from a brand they love after only one bad experience, according to a Pricewaterhouse Coopers LLC customer experience survey. On the opposite side of the scale, companies saw a 30 to 50% increase in operational efficiency, a 20% to 30% rise in employee engagement, and a 10% to 30% boost in customer satisfaction after transforming their development pipelines, according to management consulting firm McKinsey & Co. Inc.

Organizations with in-house information technology teams have traditionally built out their own deployment capabilities. But custom solutions aren’t built to deal with the complexity of modern dispersed environments, according to Ian Delahorne (pictured, right), staff site reliability engineer at Patreon Inc. “It was very hard for us, with our custom-built deploy tooling, to be able to easily deploy fast and to roll back if things went wrong,” he said.

Delahorne and Andrew Backes (left), head of engineering at Armory Inc., spoke with Lisa Martin, host of theCUBE, SiliconANGLE Media’s livestreaming studio, during the AWS Startup Showcase: Open Cloud Innovations event. They discussed enterprise-scale CI/CD pipeline deployment and how Patreon streamlined Spinnaker deployment with help from Armory. (* Disclosure below.)

Custom-built tooling can’t handle enterprise-scale deployment

The trouble began when Patreon had an outage that slowed down its payment processing systems. More than an operational oops, this was a mission-critical issue. Patreon facilitates funding for artists and creators, and over 200,000 rely on the platform for income.

“Patreon is a membership platform for creators to be able to connect with their fans and for fans to be able to subscribe to their favorite creators,” Delahorne said, describing how Patreon helps creators earn a living by being connected straight to their audience.

Delahorne knew that open-source cloud-native continuous delivery tool Spinnaker could solve the issue, but he was one of just two site reliability engineers on staff. “Spinnaker is a very complex product. I didn’t have the engineering bandwidth to be able to set up, deploy, manage it,” he said.

With fortuitous timing, Armory’s marketing department had contacted Delahorne just the week before the outage. And the company’s promise to “make software delivery, continuous, collaborative, scalable and safe” was fresh in his mind.

The solution matched Patreon’s need, and a partnership was born.

How Armory adds value on top of Spinnaker

Alongside providing a platform for safe, speedy deployment, Armory provides Patreon with a policy engine module that restricts who can deploy where for compliance issues. Patreon also uses Armory’s pipelines as code module for being able to build reusable deploy pipelines so that software engineers can easily integrate Spinnaker without having to know a lot about the tool, according to Delahorne.

“Take this pipeline module and add your variables into it, and you’ll be off to the races deploying,” he said. One of the big areas Armory focuses on is how it can provide building blocks on top of open-source Spinnaker “that sites can use to tailor the solution to their needs.”

Even more added value comes from Armory’s managed products, according to Delahorne. “They have a team that’s managing our Spinnaker installation, helping us with upgrades, helping us with the issues. All that stuff that unlocks us to be able to focus on building up our creators,” he said.

Before Patreon partnered with Armory, regular deploys would take somewhere around 45 minutes each. And events that required more instances, such as the first-of-month payment processing, would take even longer. The company cut that timeframe down to 16 to 20 minutes per deploy, even when deploying to a few hundred hosts.

“That helps us bring features to creators and fans a lot faster,” Delahorne said. Plus, he rests easier knowing that “we have a secure way to roll back if needed.”

Enterprise-scale explained

“Enterprise scale” is a term that is rarely comprehensively defined. Often when companies think enterprise-scale, they only think about the volume of infrastructure or the volume of software that’s running at any given time, according to Backes. But there’s more to it than how many EC2 instances or containers a company is running.

The velocity of deployment, as in how much time it takes to get features out to customers, is another factor, as are stability and reliability and site-specific technologies that need integrating with a company’s existing deployment tooling. And there’s also the added complication of multiple target deployment.

“It isn’t always just EC2 … it’s going to be ECS, Lambda. All of these workloads are out there running,” Backes said. “The amount of multiple targets that are being deployed to at any given moment just keep increasing.”

These are all concerns around enterprise scale that companies need to address, according to Backes.

“So one of the key things we’re doing is building ahead of that, making sure that our features are enabling users to hit deployment scales they’ve never seen or imagined before,” he said.

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the AWS Startup Showcase: Open Cloud Innovations event. (* Disclosure: Armory sponsored this segment of theCUBE. Neither Armory nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE