Find out how Lessonly distributed the deployment workload to its developers.
Lessonly is a powerfully simple training software that helps teams learn, practice, and do better work. They enable busy teams to get on the same page, stay ahead of change, and deliver amazing experiences to customers and prospects. Learn more at Lessonly.com.
Lessonly hosted its applications in Heroku, and for a time, life was good. Heroku provided deployment tools that developers used to deploy their services. But as Lessonly grew, site reliability engineers Stephen Gregory and Tyler Henrichs encountered issues. Heroku lacked infrastructure flexibility and governance features, and Heroku’s pricing structure exponentially increased the cost of hosting as the number of applications grew.
Stephen and Tyler were spending months’ worth of work to maintain and troubleshoot Heroku issues. The duo decided it was time to find a better solution. A solution that would provide advanced RBAC capability, more audit trail tracking, and help foster a “self-service” deployment culture with developers.
We wanted to create a deployment culture where developers felt responsible for their own services. - Tyler Henrichs | SRE | Lessonly
In the face of increasing Heroku expenses, Lessonly decided to migrate to AWS. Stephen and Tyler needed to figure out how they were going to deploy the new AWS applications.
The team began experimenting with Jenkins and Spinnaker-based pipelines. These Frankenstein pipelines used Jenkins to build artifacts and Spinnaker to deploy artifacts to production. During their experimentation, Tyler and Stephen found It took a full week to onboard a single service. Tyler and Stephen also estimated it would take a full 6 months to build a deployment solution using these tools, and even after 6 months, they would still lack RBAC and audit trails. They both decided delivering a lackluster solution wasn’t an option.
If we tried to build our own deployment solution, we would have degraded our developers’ experience. We weren’t about to do that. - Stephen Gregory | Principal SRE | Lessonly
After it was clear the Jenkins-Spinnaker combo wouldn’t cut it, Lessonly turned to Harness. Harness gave developers deployment ownership over their own services and came out of the box with advanced RBAC and audit capabilities. Developers at Lessonly were surprised how easy it was to pick up.
Developers thought they were going to have to learn AWS and Kubernetes, but in reality, they just had to define how the code was shipped. Took them a day. - Stephen Gregory | Principal SRE | Lessonly
Instead of taking a full week to onboard a new service, now it could be done in minutes. Stephen and Tyler no longer had to worry about the maintenance time. With their newfound free time, Stephen and Tyler directed more resources into Lessonly’s application as opposed to working on getting said application into production.
The best part of moving to Harness? Hardly any of the developers even noticed the shift.
We moved systems and changed the way everything works. No one noticed. That’s the biggest compliment. - Stephen Gregory | Principal SRE | Lessonly
UWM leverages Harness for self-service Kubernetes deployments and infrastructure creation, reducing deployment time from hours to minutes!
To Tyler Tech, feature flags were a natural extension of CI/CD. Learn why they chose - and trusted - Harness to provide that capability.
Learn how Lessonly by Seismic went from toil with Jenkins to peace of mind with Harness CIE.