About

Casting Networks, LLC is a cutting-edge entertainment technology company that helps performers find great roles and industry professionals find great talent.

Migrating to Kubernetes

Jeremy Malara, Director of DevOps at Casting Networks hit roadblocks while migrating Casting Networks legacy monolith applications to Kubernetes microservices. The purpose of this initiative was to increase developer velocity, cut costs and increase application stability. However, as this initiative kicked off it was soon apparent that the current deployment pipelines in place weren’t suited well for this migration.

On-boarding new Kubernetes microservices quickly became a bottleneck. Casting Network’s Jenkins pipelines “were made of hundreds of thousands of lines of bash script”. This complexity meant it “took an entire day to onboard a single microservice.” said Jeremy.

It took an entire day to onboard one microservice with Jenkins then I realized I had 20+ more to do. With Harness it took 10 minutes. - Jeremy Malara | Director of DevOps | Casting Networks

Knowledge silos and tribal knowledge created another migration barrier. Development teams at Casting Networks lacked experience and skills in building microservice pipelines using scripts and Jenkins pipelines. Jeremy’s team of three SRE’s were the only people capable of conducting this migration.

Jeremy understood that his approach with Jenkins Pipelines was brittle and not scalable, and thus, would hinder his team’s ability to complete their Kubernetes migration. This prompted him to evaluate other Continuous Delivery tools he could use to standardize Casting Network’s deployment process for Kubernetes and Microservices.

Standardizing and Scaling Microservice Pipelines

After a thorough evaluation of commercial and open-source tools, Jeremy soon realized that Harness Continuous Delivery as-a-Service could transform his Kubernetes migration.

Harness pipeline templates would enable standard and repeatable deployment steps, removing the need for developers or a dedicated team of SREs to script each pipeline. Each developer would have the ability to instantiate a pipeline template, and trigger deployments themselves using a self-service model. The simplicity of the Harness platform would help drive adoption and allow teams to easily onboard their microservices on their own creating fewer bottlenecks and increasing developer velocity.

Once Harness deployment pipeline templates were created, Harness Continuous Verification could be embedded into these pipelines to automate QA and canary deployments so code could be pushed straight to production with no human intervention. Casting Network’s engineers “wouldn’t even think about deployments, they would just let Harness take care of the work.”

In addition, a centralized secret store within Harness allowed teams to manage and reference secrets across their deployment pipelines in a secure and scalable manner while ensuring compliance.

Jeremy was convinced he needed Harness to jump start his Kubernetes migration initiative.

Reducing Deployment and Onboarding time by 95%

Jeremy’s team did their first deployment using Harness on Day 1. The time it took to deploy an artifact was reduced by 95% from 12 minutes to 30 seconds. The time it took to onboard a single microservice was reduced 97% from 1 day to just 10 minutes. As a result of implementing Continuous Verification, change failure rate was reduced by 90% from 20% to just 2%.

Most importantly, “Without Harness, Casting Networks would’ve been 4 months behind on their migration.” Harness helped them achieve their business goal faster.

We would’ve been 4 months behind on our Kubernetes migration without Harness. - Jeremy Malara | Director of DevOps | Casting Networks