KCL Operator provides cluster integration, allowing you to use Access Webhook to generate, mutate, or validate resources based on KCL configuration when apply resources to the cluster. Webhook will capture creation, application, and editing operations, and execute KCLRun on the configuration associated with each operation, and the KCL programming language can be used to
- Add labels or annotations based on a condition.
- Inject a sidecar container in all KRM resources that contain a
PodTemplate
. - Validate all KRM resources using KCL schema.
- Use an abstract model to generate KRM resources.
- Install Go 1.23+
- Install Kubectl and Kustomize
- Install Operator SDK
- Prepare a Kubernetes Cluster e.g., K3d
Run make help
to get the help.
- Deploy the KCL Operator.
make deploy
Use the following command to watch and wait for the pod status is Running
.
kubectl get po
- Deploy the KCL source
kubectl apply -f- << EOF
apiVersion: krm.kcl.dev/v1alpha1
kind: KCLRun
metadata:
name: set-annotation
spec:
params:
annotations:
managed-by: kcl-operator
source: oci://ghcr.io/kcl-lang/set-annotation
EOF
- Validate the mutation result by creating a nginx Pod YAML.
kubectl apply -f- << EOF
apiVersion: v1
kind: Pod
metadata:
name: nginx
annotations:
app: nginx
spec:
containers:
- name: nginx
image: nginx:1.14.2
ports:
- containerPort: 80
EOF
kubectl get po nginx -o yaml | grep kcl-operator
The output is
managed-by: kcl-operator
We can find the annotation managed-by=kcl-operator
is added on the pod.
Here's what you can do in the KCL script:
- Read resources from
option("resource_list")
. Theoption("resource_list")
complies with the KRM Functions Specification. You can read the input resources fromoption("items")
and theparams
fromoption("params")
. - Return an error using
assert {condition}, {error_message}
. - Log variable values using the function
print(variable)
and it will be output to the stdout of the pod. - Read the PATH variables. e.g.
option("PATH")
. - Read the environment variables. e.g.
option("env")
.
A KRM YAML list means that each document must have an apiVersion
, kind
through the items
field or a single YAML output.
- Using the
items
field
apiVersion: krm.kcl.dev/v1alpha1
kind: KCLRun
metadata:
name: basic
spec:
source: |
items = [{
apiVersion: "v1"
kind: "Foo"
metadata.name = "foo"
}, {
apiVersion: "v1"
kind: "Bar"
metadata.name = "bar"
}]
- Single YAML output
apiVersion: krm.kcl.dev/v1alpha1
kind: KCLRun
metadata:
name: basic
spec:
source: |
{
apiVersion: "v1"
kind: "Foo"
metadata.name = "foo"
}
You can directly use KCL standard libraries such as regex.match
, math.log
.
See here to study more features of KCL.
See here for more examples.