08-21, 14:15–14:25 (America/Halifax), Scotiabank Lecture Hall
In large organizations using Kubernetes with multiple clusters, each minor change in the Kubernetes manifest is critical. It must be reviewed and audited and requires an easy rollback mechanism. Not to mention it must be scalable.
This talk will include a short explanation of bottlenecks if we manage Kubernetes in traditional CD (Continuous Deployment) pipelines and how GitOps methodology using ArgoCD helps to overcome them.
ArgoCD is a simple and powerful GitOps tool to manage Kubernetes. To explain in short, we write our Kubernetes manifests and store them in Git cloud (ie. Github/Gitlab), then we install ArgoCD in our Kubernetes cluster(s). Now we tell ArgoCD to apply and keep watch on the changes of Git inside the cluster it is in and we never touch the cluster again to let ArgoCD manage! Whatever changes we do, we only do in Git.
In the talk, I will include a hypothetical scenario in large organizations having multiple clusters in different regions with multiple environments (production, staging, qa etc.). I will then explain the solution using traditional pipelines and then ArgoCD. How the latter can scale with multiple teams, how it solves the common security concerns compared to the traditional approach and most importantly scaling as the team and number of applications inside Kubernetes increases.
I will also prepare a short live demo of how everything works end to end.
I'm currently pursuing a Master's in Applied Computer Science at Dalhousie University. I've made contributions to open-source projects like Spotify, Elastic, and Argo. Alongside my studies, I work part-time as a Software Engineer at Akuity, a cloud-native company. This role has familiarized me with standard practices for orchestrating containers on Kubernetes.