Can we use doctl as a cli to create and destroy kubernetes clusters?

doctl has supported Kubernetes since version 1.12.0; find the most recent release here.

$ doctl kubernetes cluster create --help
create a cluster

Usage:
  doctl kubernetes cluster create <name> [flags]

Aliases:
  create, c

Flags:
      --auto-upgrade                whether to enable auto-upgrade for the cluster
      --count int                   number of nodes in the default node pool (incompatible with --node-pool) (default 3)
  -h, --help                        help for create
      --maintenance-window string   maintenance window to be set to the cluster. Syntax is in the format: 'day=HH:MM', where time is in UTC time zone. Day can be one of: ['any', 'monday', 'tuesday', 'wednesday', 'thursday', 'friday', 'saturday', 'sunday'] (default "any=00:00")
      --node-pool strings           cluster node pools, can be repeated to create multiple node pools at once (incompatible with --size and --count)
                                    format is in the form "name=your-name;size=size_slug;count=5;tag=tag1;tag=tag2" where:
                                    	- name:   name of the node pool, must be unique in the cluster
                                    	- size:   size for the nodes in the node pool, possible values: see "doctl k8s options sizes".
                                    	- count:  number of nodes in the node pool.
                                    	- tag:    tags to apply to the node pool, repeat to add multiple tags at once.
      --region string               cluster region, possible values: see "doctl k8s options regions" (required) (default "nyc1")
      --set-current-context         whether to set the current kubectl context to that of the new cluster (default true)
      --size string                 size of nodes in the default node pool (incompatible with --node-pool), possible values: see "doctl k8s options sizes". (default "s-1vcpu-2gb")
      --tag strings                 tags to apply to the cluster, repeat to add multiple tags at once
      --update-kubeconfig           whether to add the created cluster to your kubeconfig (default true)
      --version string              cluster version, possible values: see "doctl k8s options versions" (default "latest")
      --wait                        whether to wait for the created cluster to become running (default true)

Global Flags:
  -t, --access-token string   API V2 Access Token
  -u, --api-url string        Override default API V2 endpoint
  -c, --config string         config file (default is $HOME/.config/doctl/config.yaml)
      --context string        authentication context name
  -o, --output string         output format [text|json] (default "text")
      --trace                 trace api access
  -v, --verbose               verbose output

Creating a cluster using the default options is as easy as running:

doctl k8s cluster create example-cluster-01

Here’s a more complete example of creating a cluster while specifying non-default options:

$ doctl k8s cluster create \
   --region lon1 \
   --version 1.14.2-do.0 \
   --tag demo \
   --size s-2vcpu-4gb \
   --count 5 \
   --maintenance-window="tuesday=20:00" \
   --auto-upgrade \
   example-cluster-02

Notice: cluster is provisioning, waiting for cluster to be running
....................................................
Notice: cluster created, fetching credentials
Notice: adding cluster credentials to kubeconfig file found in "/home/asb/.kube/config"
Notice: setting current-context to do-lon1-example-cluster-02
ID                                      Name                  Region    Version        Auto Upgrade    Status     Node Pools
a5b355ab-0406-4e36-9ec8-3d9c70b4525e    example-cluster-02    lon1      1.14.2-do.0    true            running    example-cluster-02-default-pool

By default, the cluster’s kubeconfig will be saved locally when you create it. You can grab the kubeconfig for an existing cluster using:

$ doctl k8s cluster kubeconfig save <cluster-id|cluster-name>

If you have any feedback, please feel free to open an issue on GitHub.