website: Add deployment operator concept (#508)

Add deployment operator document.

Signed-off-by: Kfir Toledo <[email protected]>

website/content/en/docs/concepts/_index.md

@@ -1,5 +1,5 @@
 ---
 title: Core Concepts
-description: Core Concepts and Tasks
+description: Core Concepts of the ClusterLink system.
 weight: 30
 ---

website/content/en/docs/concepts/peers.md

@@ -9,7 +9,7 @@ A *Peer* represents a location, such as a Kubernetes cluster, participating in a
  it wishes to share with other peers. A peer is managed by a peer administrator,
  which is responsible for running the ClusterLink control and data planes. The
  administrator will typically deploy the ClusterLink components by configuring
- the [deployment CR]({{< ref "users#deploy-cr-instance" >}}). They may also wish to provide
+ the [deployment CR]({{< ref "operator#deploy-cr-instance" >}}). They may also wish to provide
  (often) coarse-grained access policies in accordance with high level corporate
  policies (e.g., "production peers should only communicate with other production peers").
 

website/content/en/docs/getting-started/users.md

@@ -58,64 +58,21 @@ To set up ClusterLink on a Kubernetes cluster, follow these steps:
    ```
 
    This command will deploy the ClusterLink operator on the `clusterlink-operator` namespace
-    and convert the peer certificates to secrets in this namespace.
+    and convert the peer certificates to secrets in the namespace where ClusterLink components will be installed.
+    By default, the `clusterlink-system` namespace is used.
     The command assumes that `kubectl` is set to the correct peer (K8s cluster)
     and that the certificates were created by running the previous command on the same working directory.
     If they were not, use the flag `--path <path>` for pointing to the working directory
     that was used in the previous command.
     The `--fabric` option is optional, and by default, "default_fabric" will be used.
     The `--autostart` option will deploy the ClusterLink components in the `clusterlink-system` namespace,
     and enable ClusterLink in the cluster.
+    For more details and deployment configuration see [ClusterLink deployment operator]({{< ref "operator" >}}).
 
 {{< notice note >}}
 To deploy ClusterLink on another cluster, repeat steps 2-3 in a console with access to the cluster.
 {{< /notice >}}
 
-### Additional configurations
-
-* Setting ClusterLink namespace:
-
-   ```sh
-   clusterlink deploy peer --autostart --name <peer_name> --fabric <fabric_name> \
-         --namespace <namespace>
-   ```
-
-   The `--namespace` parameter determines the namespace where the ClusterLink components are deployed.
-    Note that you must set `--autostart`, and the namespace should already exist.
-
-* Setting ClusterLink ingress type:
-
-   ```sh
-   clusterlink deploy peer --autostart --name <peer_name> --fabric <fabric_name> \
-         --ingress <ingress_type>
-   ```
-
-   The `--ingress` parameter controls the ClusterLink ingress type, with `LoadBalancer` being the default.
-    If you're using a local kind cluster, replace `<ingress_type>` with `NodePort`.
-    For a cluster running on cloud, use `LoadBalancer` or leave ingress type unspecified.
-    Note that `--autostart` must be set when configuring ingress type.
-
-* {{< anchor deploy-cr-instance >}}Full configuration setting using a ClusterLink K8s custom resource object:
-
-    ```yaml
-    kubectl apply -f - <<EOF
-    apiVersion: clusterlink.net/v1alpha1
-    kind: ClusterLink
-    metadata:
-        namespace: clusterlink-operator
-        name: <peer_name>
-    spec:
-        ingress:
-            type: <ingress_type>
-        namespace: clusterlink-system
-    EOF
-    ```
-
-    After the operator is installed, you can deploy ClusterLink by applying a ClusterLink object as shown above.
-    The ClusterLink operator will create the ClusterLink gateway components in the `clusterlink-system` namespace.
-    For more details and information about the ClusterLink custom resource,
-    refer to the [operator documentation](https://github.com/clusterlink-net/clusterlink/blob/main/design-proposals/project-deployment.md#clusterlink-crd).
-
 ## Try it out
 
 Check out the [ClusterLink Tutorials]({{< ref "tutorials" >}}) for setting up

website/content/en/docs/tasks/_index.md

@@ -0,0 +1,5 @@
+---
+title: Tasks
+description: How to do single specific targeted activities with ClusterLink.
+weight: 35
+---

website/content/en/docs/tasks/operator.md

@@ -0,0 +1,109 @@
+---
+title: Deployment Operator
+description: Usage and configuration of the ClusterLink deployment operator.
+weight: 50
+---
+The ClusterLink deployment operator allows easy deployment of ClusterLink to a K8s cluster.
+The preferred deployment approach involves utilizing the ClusterLink CLI,
+which automatically deploys both the ClusterLink operator and ClusterLink components.
+However, it's important to note that ClusterLink deployment necessitates peer certificates for proper functioning.
+Detailed instructions for creating these peer certificates can be found in the [getting started]({{< ref "users#Setup">}}).
+
+## The common use-case
+
+The common use case for deploying ClusterLink on a cloud based K8s cluster (i.e., EKS, GKE, IKS, etc.) is using the CLI command:
+
+```sh
+clusterlink deploy peer --autostart --name <peer_name> --fabric <fabric_name>
+```
+
+The command assumes that `kubectl` is configured to access the correct peer (K8s cluster)
+and that certificates files are placed in the current working directory.
+If they are not, use the flag `--path <path>` to reference the directory where certificate files are stored.
+The command deploys the ClusterLink operator in the `clusterlink-operator` namespace and converts
+the peer certificates to secrets in the `clusterlink-system` namespace, where ClusterLink components will be installed.
+By default, these components are deployed in the `clusterlink-system` namespace.
+The `--autostart` option will create a ClusterLink K8s custom resource object and deploy it to the operator.
+The operator will then create the ClusterLink components in the `clusterlink-system` namespace and enable ClusterLink in the cluster.
+Additionally, a `LoadBalancer` service is created to allow cross-cluster connectivity using ClusterLink.
+
+## Deployment for Kind environment
+
+To deploy ClusterLink in a local environment like Kind, you can use the following command:
+
+```sh
+clusterlink deploy peer --autostart --name <peer_name> --fabric <fabric_name> --ingress=NodePort --ingress-port=30443
+```
+
+The Kind environment doesn't allocate an external IP to the `LoadBalancer` service by default.
+In this case, we will use a `NodePort` service to establish multi-cluster connectivity using ClusterLink.
+Alternatively, you can install MetalLB to add a Load Balancer implementation to the Kind cluster. See instructions
+[here](https://kind.sigs.k8s.io/docs/user/loadbalancer/).
+The port flag is optional, and by default, ClusterLink will use any allocated NodePort that the Kind cluster provides.
+However, it is more convenient to use a fixed setting NodePort for peer configuration, as demonstrated in the [ClusterLink Tutorials]({{< ref "tutorials">}}).
+
+
+## Deployment of specific version
+
+To deploy a specific ClusterLink image version use the `tag` flag:
+
+```sh
+clusterlink deploy peer --autostart --name <peer_name> --fabric <fabric_name> --tag <version_tag>
+```
+
+The `tag` flag will change the tag version in the ClusterLink K8s custom resource object that will be deployed to the operator.
+
+## Deployment using manually defined ClusterLink custom resource
+
+The deployment process can be split into two steps:
+
+1. Deploy only ClusterLink operator:
+
+    ```sh
+    clusterlink deploy peer ---name <peer_name> --fabric <fabric_name>
+    ```
+
+    This command will deploy only the ClusterLink operator and the certificate's secrets as described in the [common use case]({{< ref "operator#the-common-use-case" >}}).
+
+1. {{< anchor deploy-cr-instance >}} Deploy a ClusterLink K8s custom resource object:
+
+    ```yaml
+    kubectl apply -f - <<EOF
+    apiVersion: clusterlink.net/v1alpha1
+    kind: ClusterLink
+    metadata:
+        namespace: clusterlink-operator
+        name: <peer_name>
+    spec:
+        ingress:
+            type: <ingress_type>
+        dataplane:
+            type: envoy
+            replicas: 1
+        logLevel: info
+        namespace: clusterlink-system
+    EOF
+    ```
+
+## Additional deployment configurations
+
+The `deploy peer` command has the following flags:
+
+- **namespace:** This field determines the namespace where the ClusterLink components are deployed.
+ By default, it uses `clusterlink-system`, which is created by the `clusterlink deploy peer` command.
+ If a different namespace is desired, that namespace must already exist.
+- **dataplane:** This field determines the type of ClusterLink dataplane, with supported values `go` or `envoy`. By default, it uses `envoy`.
+- **dataplane-replicas:** This field determines the number of ClusterLink dataplane replicas. By default, it uses 1.
+- **ingress:** This field determines the type of ingress service to expose ClusterLink deployment,
+  with supported values: `LoadBalancer`, `NodePort`, or `None`. By default, it uses `LoadBalancer`.
+- **ingress-port:** This field determines the port number of the external service.
+  By default, it uses port `443` for the `LoadBalancer` ingress type.
+  For the `NodePort` ingress type, the port number will be allocated by Kubernetes.
+  In case the user changes the default value, it is the user's responsibility to ensure the port number is valid and available for use.
+- **log-level:** This field determines the severity log level for all the components (controlplane and dataplane).
+  By default, it uses `info` log level.
+- **container-registry:** This field determines the container registry to pull the project images.
+  By default, it uses `ghcr.io/clusterlink-net`.
+- **tag:** This field determines the version of project images to pull. By default, it uses the `latest` version.
+
+All the flags are mapped to corresponding field in the ClusterLink custom resource.