As more and more tech teams lean into DevOps, it’s become increasingly important to support those processes with a robust set of tools. But given the growing number of DevOps tools available, it can be tough to know which ones are right for the job.
Skillsoft doesn’t see it that way.
By categorizing their toolchain into four key technologies, engineering teams are able to fully encapsulate Skillsoft’s infrastructure, configuration and code as versionable constructs, Senior Director of Software Engineering Anil D’silva said.
Built In Boston sat down with D’silva to uncover what other tech is making Skillsoft’s processes easier and how his teams are encouraged to try new tools that best suit their needs.
Skillsoft uses digital corporate learning solutions to help companies share institutional knowledge across their teams. The edtech business offers safety training, leadership development and guidance in certification for tech and development departments. D’silva said that by investing in the right tools, his teams are now achieving results that weren’t possible previously.
Give us a brief glimpse into your DevOps toolchain. What are a few of your favorite tools your team is currently using?
The DevOps toolchain at Skillsoft is based on four key technologies: Terraform for AWS core-infrastructure, Ansible for application deployments, Jenkins for deployment orchestration and Git as our source management system. We use a Gitflow model of deployments that deploy our 100+ microservices to Kubernetes clusters using custom-built operators. This toolset enables us to fully encapsulate our infrastructure, configuration and code as version able constructs in our source repositories. Some of our favorite tools continue to be Jenkins for managing how our CI/CD processes are run, and the Kubernetes APIs to manage our services in various environments.
What were some of the key considerations when evaluating and choosing the tools your team would use?
At Skillsoft, development teams are encouraged to use tools that get the task accomplished for their team. For example, in the last year, we have introduced a number of tools like Cerebro and Kafdrop onto our platform as suggested by the teams that were using Elastic and Kafka respectively. If a tool becomes widely adopted, the DevOps teams take on the ownership and standardization of that tool, hereby guaranteeing that only the tools actively being used by the wider community are maintained and managed.
To evaluate engineering-wide tools, we typically do a fair amount of research and due diligence on the landscape of the services available to solve that pain point. We will invite best-of-breed tool vendors to help us proof of concept and determine if the tool meets our needs, and then we build a cost-benefit analysis of the competing tools versus us building the functionality ourselves. Armed with this information, we will make a data-driven decision on the best way forward. This is exactly the path we are talking about right now as we look to reinvent our deployment pipelines.
“At Skillsoft, development teams are encouraged to use tools that get the task accomplished for their team.”
How has your DevOps toolchain evolved over time, and why?
Since mid-2020, the DevOps and operations teams at Skillsoft have invested heavily in building a common model and toolchain for all environments to follow in terms of configuration and infrastructure management. We have recently finished this effort, with all of our development to production environments being exactly the same in terms of configuration and configurability. This has immediately allowed us to do things we could have never done before, like testing out and automating database version upgrades in development.
The consistency of environments also allows our configuration to be located in one place versus the variety of repositories. This quality alone has saved countless hours debugging issues that crop up when configuration dependencies are spread out over various places.
