Run Symphony on Kubernetes

This tutorial shows you how to deploy Symphony to a Kubernetes cluster and use kubectl to interact with Symphony API in addition to REST API.

Part 1: Deploy Symphony to a Kubernetes cluster

Install Maestro and deploy Symphony to a Kubernetes cluster, to which your kubectl is currently pointing:

    wget -q https://raw.githubusercontent.com/eclipse-symphony/symphony/master/cli/install/install.sh -O - | /bin/bash
    maestro up

    powershell -Command "iwr -useb https://raw.githubusercontent.com/eclipse-symphony/symphony/master/cli/install/install.ps1 | iex"
    maestro up

You should see outputs like:

Checking kubectl ...found
Checking kubectl context ...OK
Checking Kubernetes connection ...OK
Installing cert manager ...done
Checking Helm ...found
Checking Symphony API (Symphony) ...done
Checking namespace existence ...done
Deploying Symphony API (Symphony), chart: oci://ghcr.io/eclipse-symphony/helm/symphony, installServiceExt: true, with MQTT broker: false
Deploying Symphony API (Symphony) ...done
Checking public Symphony API address ...
Symphony API: http://<Symphony REST API IP>:8080/v1alpha2/greetings

Done!

Check Symphony’s pod using kubectl:

kubectl get pods

You shoulde see a list of pods running:

NAME                                   READY   STATUS  
symphony-api-...                       1/1     Running 
symphony-cert-manager-...              1/1     Running 
symphony-cert-manager-cainjector-...   1/1     Running 
symphony-cert-manager-webhook-...      1/1     Running 
symphony-controller-manager-...        2/2     Running 
symphony-redis-...                     1/1     Running 
symphony-zipkin-...                    1/1     Running

Part 2: Deploying a simple service

Use maestro to deploy the hello-k8s sample, which deploys a Prometheus service on Kubernetes:

maestro samples run hello-k8s

You should see outputs like this:

Creating target sample-k8s-target ... done
Creating solution sample-prometheus-server ... done
Creating solutionversion sample-prometheus-server-v-v1 ... done
Creating instance sample-prometheus-instance ... done
... checking post-command condition
... checking post-command condition
⣯  NOTE: Navigate to http://<Prometheus service public IP>:9090/ to access the Promethus portal (it may take a few minutes for the LoadBalancer to be provisioned)

Navigae to the above URL to visit Prometheus portal.

The sample creates a sample-k8s-scope namespace and deploys several resources. Use kubectl to examine them:

kubectl get all -n sample-k8s-scope

You should see something like:

AME                                         READY   STATUS    RESTARTS   AGE
pod/sample-prometheus-instance-...          1/1     Running   0          38s

NAME                                        TYPE           CLUSTER-IP     EXTERNAL-IP       PORT(S)          AGE
service/sample-prometheus-instance          LoadBalancer   10.0.214.131   <Public IP>       9090:32171/TCP   33s

NAME                                         READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/sample-prometheus-instance   1/1     1            1           38s

NAME                                         DESIRED   CURRENT   READY   AGE
replicaset.apps/sample-prometheus-...        1         1         1       38s

You should see something like:

... IMAGE           ... CREATED          STATUS             PORTS                       NAMES
... prom/prometheus ... 18 seconds ago   Up 17 seconds      0.0.0.0:9090->9090/tcp      sample-prometheus-server

Part 3: Examine Symphony objects

The sample deploys several Symphony objects:

A sample-k8s-target Target object that represents a deployment target
A sample-prometheus-server Solution object with a sample-prometheus-server-v-v1 version that represent the software being managed
A sample-prometheus-instance Instance object that describes a deployment of the software to the target

NOTE: See Orchestration Model for more details.

You can examine these objects using kubectl:

kubectl get target
kubectl get solution
kubectl get solutionversion
kubectl get instance

Alternatively, you can use maestro to query the same objects. maestro goes through the REST API, but you should see identical objects via either channel:
```
maestro get target
maestro get solution
maestro get solutionversion
maestro get instance
```
maestro also support direct JSONPath queries, for example:
```
maestro get instance --json-path "@.spec.target"
```
This displays the Target name associated with the Instance:
```
NAME
sample-k8s-target
```

Part 3: Clean up

Remove the Promethus service by removing the Instance object:
```
kubectl delete instance sample-prometheus-instance
```
Once the Instance object is removed, Symphony state reconcilation removes what it represents. You can check with kubecctl:
```
kubectl get all -n sample-k8s-scope
```
This should return an empty resource list:
```
No resources found in sample-k8s-scope namespace.
```
NOTE: Symphony Kubernetes provider doesn’t remove namespaces

(Optional) Remove other Symphony objects:

kubectl delete solutionversion sample-prometheus-server-v-v1
kubectl delete solution sample-prometheus-server
kubectl delete target sample-k8s-target

Or, remove the namespece altogether:

kubectl delete ns sample-k8s-scope