How Cloud Academy sped up data modeling by 5x and decreased analytics downtime by 90%

The Background

Cloud Academy is a SaaS startup based in San Francisco, CA. Their e-learning platform aims to give tech workers a personalized upskilling experience in technical subjects such as cloud technology, DevOps, and software development.

“Personalized” is somewhat of an understatement: Cloud Academy’s training resources and assessments are customized precisely to the user’s trajectory, role, and skills. In addition, the analytics and personalization granularity of the platform offers organizations robust insights into their teams’ proficiency. With these insights, companies can boost employee retention and accelerate company operations.

The Challenge

Cloud Academy was searching for a modern analytics platform to improve its existing enterprise analytics offering.

Alessandro Lollo, Senior Data Engineer, explains that the team needed to deliver a seamless, highly available embedded analytics experience based on CI/CD best practices that also allowed planned outages for required infrastructure maintenance. They also needed to leverage their existing data warehouse and maintain high flexibility in their data modeling to serve both internal and external end users.

Before Cube, Cloud Academy used a major BI platform on which their enterprise embedded analytics UI was based; internal stakeholders used the BI platform directly. However, the team needed a solution that would allow for greater flexibility when it came to effective data modeling, security context orchestration, and caching.

The Requirements

Cloud Academy was looking for a solution that could serve as the basis of their embedded analytics while also offering:

• Compatibility with existing data warehouse and datasets
• Highly flexible data modeling
• Highly flexible security context orchestration and data visibility configuration
• Fully managed platform with zero downtime and multi-deployment
• Easy integration with existing CI/CD process
• Highly flexible, powerful, and fully managed caching
• SQL API

“With Cube, we’ve been able to speed up time to release a new data model to production by 5x and decrease analytics downtime by 90%. And, since the Cube team is so responsive, collaborative, and fast to deliver, we benefit from new features very frequently.” —Alessandro Lollo, Senior Data Engineer

The Solution

Lollo’s team found Cube in mid-2021 and chose it for its highly flexible caching layer and fully managed, Git-integrated, and zero-downtime deployments.

It took ten engineers about a month to onboard and integrate Cube with Cloud Academy’s existing CI/CD platform. After these initial processes, they delivered their first Cube-based analytics app to enterprise customers in less than three weeks.

Cube’s flexible data modeling framework made it easy for the team to use their existing data warehouse as a single source of truth, allowing them to serve consistent data models to both enterprise customers and internal stakeholders.

In addition, Cube Store’s fully managed pre-aggregation capabilities allowed them to boost the response times of complex data models that serve both internal and external use cases.

Lastly, Cube Cloud’s isolated deployment and Git integration enabled the team to implement true and efficient SDLC, while its RBAC framework has served as a component in security context orchestration.

With Cube, Cloud Academy was able to speed up releases of new data models in production by 5x while decreasing their platform’s analytics downtime by 90%.

Another surprising benefit was Cube’s Alerts feature, which provided the team with a simple yet effective mechanism to monitor deployments’ API health and pre-aggregation build failures.

The cherry on top? “Since the Cube team is responsive, collaborative, and fast to deliver, we benefit from new features very frequently,” says Lollo.

The Future

As they progress in their Cube implementation, Lollo’s team is planning to expand into incremental pre-aggregations and experiments using Cube’s SQL API joins.

Looking ahead, they’re most excited for future Cube releases including metrics layer, Cube Store’s alternative to Redis, and YAML support.

As for advice for others considering Cube Cloud, Lollo says, “When we first found Cube, we considered hosting it internally. However, we realized doing so would require quite the overhead from our DevOps team. If you plan on using Cube but don’t want to (or can’t) increase your infrastructure OPEX, then go for Cube Cloud—you won’t regret it.”

Have an exciting Cube case study and want to be featured? Drop us a line: hello@cube.dev

Interested in managed Cube Cloud for your data modeling, access control, and caching needs? Request a 1:1 here, or find us on Slack and GitHub.