Services #
We briefly touched upon services earlier in the introduction. We will now explore services a little bit deeper.
Endpoints #
When you create a service, you will be able to see that it picked up pods by looking at the endpoints. The following command should show that your service has found a list of IPs which correspond to the IPs of the pods that it is matching.
kubectl get endpointslice
Service type #
The Kubernetes Service resource can be of different types. If we don’t specify
the type it will default to ClusterIP. You can read more about other service
types in
the sevice
documentation
Exercises #
Service load balancing in action #
In this exercise we will deploy an application which simply writes a log each
time it gets a request. The source code for the application can be found in
exercises/kubernetes/services. I made it easy for you to deploy by supplying
the Deployment resource for you at
exercises/kubernetes/services/deployment.yaml
- Apply the
deployment.yamlto the cluster - Fill out the specification of the service in
exercises/kubernetes/services/service.yaml, and apply it to the cluster. - Use the
sternCLI to watch the logs of the deploymentstern deployment/http-logger. - Run
kubectl proxy. This enables us to send requests to the service from our host machine. - Send requests to the service using
curl http://localhost:8001/api/v1/namespaces/default/services/http-logger:80/proxy/ - Notice how different pods are hit in the stern log output.