Playing with the Kubernetes API
In my journey to learn the Kubernetes API, the first part is to connect to it.
You could use simple curl using credentials from $KUBECONFIG to do it, but
Kubernetes provides a simpler and better way to connect to it: kubectl proxy --port=8080.
This command will create a proxy between your localhost and your target kubernetes API
according to your current kubernetes context.
After that you could keep using curl to just communicate to the api like
curl localhost:8080/apis/apps/v1/deployments; however, you can replicate the
same with kubectl using kubectl get --raw=<uri>. For instance, the previous
curl command can be translated to kubectl get --raw=/apis/apps/v1/deployments.
Watch - Checking Kubernetes events in a resource
If you are curious about which events are happening in a resource, you can use
watch for it. For this, you need the Kubernetes object resource version which
can be used to initiate a watch against the server. The server will return all
changes(creates, deletes and updates) that occur after the supplied resourceVersion.
#Starting proxy
$kubectl proxy --port=8080
#Obtaining resourceVersion
$RESOURCE_VERSION=$(kubectl get --raw=/api/v1/namespaces/default/pods | jq -r .metadata.resourceVersion)
#Starting watch
$kubectl get --raw=/api/v1/namespaces/default/pods?watch=1&resourceVersion=${RESOURCE_VERSION}
The watch in Linux starts as a background job by default. You can check it with
jobs. After that, you can bring it to the foreground using fg [job_number].
If the client disconnects, you can restart it using the last returned resourceVersion,
or performing a new request and begin again.