We had a great turn out for our Kubernetes Webinar yesterday. Answering questions we could not get to from the audience:
- Is Harness available as a setup in (on) premises?
a. We do have connected and disconnected on-premise options.
- Is there any cost saving when we run the workload in VM vs Kubernetes? How do I ensure, my cluster is configured efficiently to optimize utilization and ensure it is cost effective?
a. From cost savings you are typically getting more density out of a container thus will get more density on the workload that Kubernetes orchestrates compared to a traditional VM. Kubernetes is a platform and there are many vendor choices where to spin up and manage your Kubernetes Cluster(s). A distributed system challange making sure clusters are right-sized based on workload types. The worker node size and capacity are important. Not sure if you are looking at public cloud vs internal options cost and usage metrics would be different.
- How is harness different from other CD platforms like Gitlab and CodeFresh?
a. The Harness Platform is designed to provide best of breed capabilities around orchestrating and decisioning on the multiple pieces required to get software into production. We leverage several methods including Machine Learning to validate and continuously validate deployments.
- What are the roles (SRE / devOps engineer) / responsibilities and skills required for setup and managing K8s?
a. The K8s ecosystem and platform continue to increase. With the container bloom, there are those who create the images [what turns into a running container]and those who administer and scale the platforms where the containers run e.g Kubernetes, Mesos, etc. From a DevOps prospective, if platform engineering falls under a DevOps title would be understanding the moving pieces e.g getting a CKA, Certified Kubernetes Administer certification. From the SRE prospective, understanding the applications and limitations and the “why” Kubernetes. SRE wise the clustering and consensus protocols that the applications being deployed requires.
- While on GCP, why would someone choose Harness over GKE?
a. Harness is not a Kubernetes provider we are an orchestrator of Continuous Delivery. Many of our customers do deploy into GKE.