Gloo was founded with a noble purpose—to empower champions to help the growth of others, at scale. The organization connects people who are struggling in their lives with the ministers, counselors, coaches, and mentors on the frontlines, who guide people through personal transformation, whether they are facing addiction, suicidal ideation, marriage troubles or work problems.

The chaos and complexity of unruly infrastructure

While trying to fulfill its purpose, Gloo faced some large technical and operational challenges.  

The platform’s infrastructure was built on top of AWS and Kubernetes, with many open-source tools stitched together in a less-than-optimal manner. This introduced a level of complexity that was eventually impossible to manage, improve, and keep up-to-date, even for the 6 SREs on the Gloo  team—a headcount that cost the company north of $1 million per year, excluding the cost of the infrastructure itself. McCurdy said, “We were getting to a spot where everybody was having to worry about the infrastructure. I have to commend the people that put our infrastructure together, but it just wasn’t serving us. It was too fragile and required too much babysitting.” And as McCurdy found out later, they had a lot of extra costs hiding in that infrastructure.

“Most developers want to write code, deploy the thing, and just have it work,” said McCurdy. “Our developers want to focus on delivering value to our customers, not messing with Helm charts, Terraform, and the very specific ops-related knowledge that often comes along with having a home-grown infrastructure.” 

This setup imposed intricate processes, which took extensive manual labor to execute—something that slowed down their product engineers and prevented the team from shipping in a timely manner. It also required a great deal of human supervision from the company’s SREs. Neither were sustainable for a company of Gloo’s size.

“The infrastructure that we had set up with Kubernetes was painful and getting in the way of our engineers. There were a lot of random build failures. We had to think about it too much. I want to not think about infrastructure. We knew there had to be a better way.”

Connected to this challenge was the threat of knowledge loss. McCurdy explained, “So you’d have a piece of the infrastructure where potentially one person knew how it all worked. And then if that person left, we were put into a less-than-desirable position. It’s the classic ‘bus factor.’”

The search for a secure platform

Because people share sensitive information through Gloo’s platform, protecting the privacy of users is the highest commitment of the company. “We place privacy and trust above all else, and do our best to make secure software as rapidly as possible, because there are literally lives on the line.” 

For this reason, McCurdy’s top priority was finding a platform-as-a-service with equal strengths in security and compliance. McCurdy looked at several solutions including Heroku, Render, Platform.sh as well as Aptible when looking for a cloud platform. Some of the companies were immediately disqualified by the security needs that Gloo required. 

Aptible was the only option that prioritized HIPAA compliance, and didn’t complicate any of our SOC 2 audits, McCurdy explained. Plus, the platform’s security, privacy and compliance focus streamlined Gloo’s compliance audits. “You know all those reports you have to fill out? With Aptible’s compliance dashboard, you just generate the report, hit the print button and go,” said McCurdy.  

The clincher was the responsiveness and support that Aptible showed. McCurdy said, “It became pretty apparent that Aptible was the only choice for what we were looking to do quite honestly. And what put it over the edge was the support. Ben and his crew were just on top of it and answered every question and addressed everything we ran into.”

The benefits of making the right choice 

McCurdy sees Aptible as a quality-of-life tool because it significantly improves his team’s developer experience. “It just does exactly what you want it to do. The fact that all I have to do is build a Docker image, ship to ECR or any other container registry, and then pull it into Aptible is a wonderful quality of life improvement  for me and my team.”

Conceptually, Aptible is not hard to understand and that makes it easy to implement and use.

“It’s simple like Heroku, but way more powerful. We didn’t have to compromise on anything, and I can’t think of a use case that Aptible wouldn’t work for.”

Product engineers no longer have to worry about infrastructure, and they’re far more productive as a result. “We like not having to tinker with infrastructure,” said McCurdy. “And that lets us focus on what’s most important: serving our users well.” Gloo also found that its concerns about losing institutional knowledge when people left were gone. “Part of the reason Aptible’s been so great is that I don’t have to worry about that,” said McCurdy.

Looking back, it’s clear that switching to Aptible was a wise investment. “Costs are down,” said McCurdy. “We have visibility into the infrastructure and previously we had a lot of phantom containers running out there. We are running our primary applications at a fraction of the cost it was costing us before - and that’s with ten times the users. It’s incredible.” It’s worth repeating that even as users went up 10x, overall costs went down.

The best part of being an Aptible customer?

“The support we get is bananas. The team is incredibly responsive to our needs. They've almost become an extension of our team in many ways,” said McCurdy. Aptible Support Engineers and Gloo developers share a common Slack channel. “Everyone’s always so eager to help. It’s truly phenomenal.”