How can I scale my google compute engine that hosts my kubernetes cluster?

Question

Description :
Ideally (i.e in a non kubernetes scenario where my compute engines is hosting my application ) a load balancer would distribute the load on multiple replicated version of compute engines. But in case when I am using just my compute engine as worker node and it has some pods deployed on it.

Question 1 :
What would happen if my worker node ( a google computer engine ) starts receiving a lot of traffic.

Question 2 : What would be the best(or atleast a better) way to scale my current solution so that it is able to manage more load and also that my load is efficiently distributed ?

Answer 1

In Kubernetes you deploy applications as pods. You can deploy multiple replicas of pods and Kubernetes will schedule it into multiple worker node VMs based on the resource requirement of the pods and available capacity on the nodes.This will provide resiliency and availability for applications. Once your workload increases you can scale the kubernetes cluster horizontally by adding more worker nodes.

You can use an ingress or L7 Loadbalancer to load balance user traffic onto the pods across different nodes. Even without those kubernetes provides L4 load balancing via kube proxy component.

Kubernetes scales to 5000 nodes. Some best practices for large cluster.