Photo by Andrew Neel on Unsplash

A Serverless Alternative to Kubernetes

Ready to Go

Containers have long been a part of the admin's daily grind for both in-house development of applications and for operating third-party products. Although Kubernetes has established itself as the market leader, the overhead required to run Kubernetes is often significant. In this article, I look at the lightweight Google Cloud Run alternative and show you how it can be set up.

As with Kubernetes, a number of requirements must be met for Cloud Run [1], which is designed to handle the deployment and management process across the entire lifecycle – from provisioning and scheduling to deletion – and significantly reduce the work involved in operating containerized applications.

Scalability is also important. Containers are typically implemented stateless; in other words, either a large number of instances are launched or they scale down to zero as a function of the load. Of course, the application must also be accessible, which requires a load balancer to distribute traffic across the back-end instances. To enable scaling, you need to monitor utilization, which typically means CPU and memory, to help avoid overprovisioning while ensuring that each container receives the resources it needs.

High availability also plays a crucial role in application operations. If a container fails, it has to restart automatically. If the underlying node has an issue, a different host must be used for the restart. Don't neglect the network, either, especially if applications need to run with high availability and scalability. Although container environments are easy to manage on a single node, manually managing larger environments involves far more effort in terms of networking; here, the container orchestrator steps in.

Cloud Run as a Kubernetes Alternative

Cloud Run is intended as a dedicated Kubernetes alternative for many use cases, as