Do You Really Need Kubernetes? Real‑World Dev Opinions and Practical Tips

The piece gathers several high‑voted Zhihu responses to the question “Do you really need Kubernetes?” and presents a balanced view of its advantages, drawbacks, and practical usage tips for developers and operations teams.

Why Some Developers Question Kubernetes

Many contributors note that Kubernetes brings significant configuration overhead and high resource consumption, making it feel excessive for small projects or low‑traffic services. One user remarks that a Spring Cloud + K8s stack often results in dozens of servers for modest daily traffic, inflating both server and development costs.

Arguments for Adopting Kubernetes

Other respondents highlight automation, scalability, and self‑healing capabilities. They point out that managed cloud clusters provide cheap master nodes, and that a single physical machine can serve as a K8s node, eliminating the need for additional virtualization layers. Experience with K8s can reduce manual SSH work, streamline CI/CD pipelines, and simplify service updates.

Typical Kubectl Commands

kubectl create deploy myapp --image=myimage

– create a deployment.

kubectl expose deploy myapp --type=LoadBalancer --target-port=80 --port=80

– expose the deployment as a service. kubectl scale deploy myapp --replicas=3 – scale the deployment for load balancing. kubectl edit ingress myapp-ingress – edit domain, routing, and whitelist settings. kubectl set image deployment/myapp myapp=image:newtag or helm upgrade … – update the application image for CI/CD.

Organizational and Team Considerations

Adopting Kubernetes requires skilled personnel. Teams must assess whether they have enough developers and ops staff to maintain the platform, handle CI/CD migration, and manage the increased operational complexity. Hiring a “K8s leader” without deep knowledge of the ecosystem can lead to gaps in container, node, and kernel management.

Furthermore, the decision should be driven by concrete problems: does the organization need automated scaling, self‑healing, or multi‑cluster management? If the answer is no, a simpler Docker‑compose or direct Docker deployment may be more appropriate.

Consensus and Recommendations

Overall, Kubernetes shines for medium to large‑scale, micro‑service‑heavy environments where its orchestration benefits outweigh the learning curve and operational overhead. For small services or teams lacking expertise, the platform can introduce more issues than it solves. The key is to evaluate business needs, team capabilities, and long‑term maintenance before committing.