CI/CD Workflow for SpringBoot Applications on Kubernetes via Skaffold

by Ashish ChoudharyDecember 14th, 2020

Too Long; Didn't Read

In this article, we'll explain how to create a CI/CD workflow for a SpringBoot Application deployed to Kubernetes using Skaffold.

Companies Mentioned

featured image - CI/CD Workflow for SpringBoot Applications on Kubernetes via Skaffold

In this article, we'll explain how to create a CI/CD workflow for a SpringBoot Application deployed to Kubernetes using Skaffold.

It is a known fact that developing applications with Kubernetes is cumbersome. That is the reason there is an ecosystem being developed around it so that developers can focus on what matters most to them i.e. writing code. On that note, in this article, I will be covering another tool Skaffold developed by Google.

So what exactly is Skaffold?

Skaffold handles the workflow for building, pushing and deploying your application.

What problem it is trying to solve?

As I said earlier developing applications with Kubernetes is not such an easy task and that’s where Skaffold comes into play as it eases the development and deployment of your applications running on Kubernetes. It manages the entire workflow and you get instant feedback while developing and deploying your application locally or on a remote Kubernetes cluster.

Following are some of the benefits of using Skaffold.

Lightweight

It’s is a client-side utility only so there is no cluster set up or anything to maintain.

Easy to share

It’s very easy to share among your team members as you only need to do the following to get started.

git clone
skaffold run

Ease of development

To get your application containerized even locally you have to do a lot of things. For your local development, you might be doing something like building, pushing, and then deploying your application on Kubernetes. And there are a different set of commands (docker blah, kubectl blah blah, etc.) and the tools you use for each of these phases of your workflow.

With Skaffold you just have one magical command skaffold run or skaffold dev and you are golden. This doesn't mean you don't have a dependency on those tools that require you to build and deploy applications to Kubernetes. It's just that development is a lot easier when you just run a single command.

OK. Enough theory let’s get started.

Anatomy of the SpringBoot Application

In this tutorial, I will be using a SpringBoot application which when accessed via /states REST endpoint shows Indian states and their capitals. This application uses an in-memory H2 database which inserts rows at the start of the application and keeps it in memory. The source code is available here.

Prerequisites

For this demo following are required to be installed.

Installing Skaffold
Installing Docker Desktop for Mac
Installing kubectl
Minikube(Optional)

I have used macOS but you are free to use any preferred OS you like or have experience working with. I have mentioned Minikube installation is optional because if you already have Docker desktop installed you can enable Kubernetes by following the below steps.

Enabling Kubernetes With Docker Desktop for Mac

Go to Preferences in the Docker menu bar, then click the Kubernetes tab. Click the checkbox to enable Kubernetes and switch the default orchestrator to Kubernetes. It will take some time to enable Kubernetes. Once it is enabled you will see something like below on your Docker menu bar. You should also change the context to docker-desktop also.
Once it is enabled you will see something like below on your Docker menu bar. You should also change the context to docker-desktop also.

Kubernetes enabled

Kubernetes Running

Context set to docker-desktop

Getting Started With Skaffold

skaffold init

To start with Skaffold you would require a skaffold.yaml file. For that, we can run the below command.

skaffold init

However, you will be greeted with the following error message.

skaffold init
one or more valid builder configuration (Dockerfile or Jib configuration) must be present to build images with skaffold; please provide at least one build config and try again or run `skaffold init --skip-build`

The error seems to be self-explanatory that Skaffold is looking for either Dockerfile or Jib configuration in your project.

Skaffold currently supports the following builders.

Docker
Jib (with --XXenableJibInit flag)
Buildpacks (with --XXenableBuildpacksInit flag)

So to resolve this error I will be adding the Jib maven plugin in pom.xml file. If you are wondering what is Jib and its usage you can read my previous article. You can enable Jib support by copy/pasting the below content to pom.xml file.

<plugin>
   <groupId>com.google.cloud.tools</groupId>
   <artifactId>jib-maven-plugin</artifactId>
   <version>2.7.0</version>
   <configuration>
      <from>
         <image>gcr.io/distroless/java:11</image>
      </from>
      <to>
         <image>registry.hub.docker.com/hiashish/indian-states</image>
      </to>
   </configuration>
</plugin>

As per official documentation.

skaffold init also recognizes Maven and Gradle projects, and will auto-suggest the jib builder. Currently jib artifact detection is disabled by default, but can be enabled using the flag --XXenableJibInit.

Now run skaffold init with --XXenableJibInit. However, it will fail again with the below error.

skaffold init --XXenableJibInit
one or more valid Kubernetes manifests are required to run skaffold

Since we have not created Kubernetes manifests(deployment, pod, service, etc.) and there is a known issue with Skaffold and to resolve this error we will have to create them manually using the below kubectl command.

Creating deployment

kubectl create deployment states --image=docker.io/hiashish/indian-states --dry-run -oyaml
apiVersion: apps/v1
kind: Deployment
metadata:
  creationTimestamp: null
  labels:
    app: states
  name: states
spec:
  replicas: 2
  selector:
    matchLabels:
      app: states
  strategy: {}
  template:
    metadata:
      creationTimestamp: null
      labels:
        app: states
    spec:
      containers:
        - image: docker.io/hiashish/indian-states
          name: indian-states
          resources: {}
status: {}

Creating service

kubectl expose deployment states --type=NodePort --port=8080 --dry-run -oyaml
apiVersion: v1
kind: Service
metadata:
  creationTimestamp: null
  labels:
    app: states
  name: states
spec:
  ports:
    - port: 8080
      protocol: TCP
      targetPort: 8080
  selector:
    app: states
  type: NodePort
status:

We have created both deployment and service now. Please make sure to copy the output of the above command to YAML files in the k8s directory.

Now run skaffold init --XXenableJibInit

skaffold init --XXenableJibInit
apiVersion: skaffold/v2beta10
kind: Config
metadata:
  name: indian-states
build:
  artifacts:
  - image: docker.io/hiashish/indian-states
    jib:
      project: com.example:indian-states
deploy:
  kubectl:
    manifests:
    - k8s/mydeployment.yaml
    - k8s/myservice.yaml
Do you want to write this configuration to skaffold.yaml? [y/n]: y
Configuration skaffold.yaml was written
You can now run [skaffold build] to build the artifacts
or [skaffold run] to build and deploy
or [skaffold dev] to enter development mode, with auto-redeploy

Finally, skaffold.yaml file is created.

skaffold dev

We have now completed the required setup to start continuous build and deployment of our Kubernetes application. Now we can simply run the below command to start our CI/CD workflow locally.

skaffold dev
Listing files to watch...
 - docker.io/hiashish/indian-states
Generating tags...
 - docker.io/hiashish/indian-states -> docker.io/hiashish/indian-states:31ff588-dirty
Checking cache...
 - docker.io/hiashish/indian-states: Found Locally
Tags used in deployment:
 - docker.io/hiashish/indian-states -> docker.io/hiashish/indian-states:43f7c470a60b876c7579ed3041b64024b774e9808851ad83b6817701d0188cc5
Starting deploy...
 - deployment.apps/states created
 - service/states created
Waiting for deployments to stabilize...
 - deployment/states is ready.
Deployments stabilized in 2.710870355s
Press Ctrl+C to exit
Watching for changes...
[indian-states] 
[indian-states]   .   ____          _            __ _ _
[indian-states]  /\\ / ___'_ __ _ _(_)_ __  __ _ \ \ \ \
[indian-states] ( ( )\___ | '_ | '_| | '_ \/ _` | \ \ \ \
[indian-states]  \\/  ___)| |_)| | | | | || (_| |  ) ) ) )
[indian-states]   '  |____| .__|_| |_|_| |_\__, | / / / /
[indian-states]  =========|_|==============|___/=/_/_/_/
[indian-states]  :: Spring Boot ::                (v2.4.0)
[indian-states] 
[indian-states] 2020-12-07 17:43:54.919  INFO 1 --- [           main] c.e.i.IndianStatesApplication            : Starting IndianStatesApplication using Java 11.0.6 on states-6f5bb746b6-9sglw with PID 1 (/app/classes started by root in /)
[indian-states] 2020-12-07 17:43:54.938  INFO 1 --- [           main] c.e.i.IndianStatesApplication            : No active profile set, falling back to default profiles: default
[indian-states] 2020-12-07 17:43:57.607  INFO 1 --- [           main] .s.d.r.c.RepositoryConfigurationDelegate : Bootstrapping Spring Data JDBC repositories in DEFAULT mode.
[indian-states] 2020-12-07 17:43:57.670  INFO 1 --- [           main] .s.d.r.c.RepositoryConfigurationDelegate : Finished Spring Data repository scanning in 35 ms. Found 0 JDBC repository interfaces.
[indian-states] 2020-12-07 17:44:00.130  INFO 1 --- [           main] o.s.b.w.embedded.tomcat.TomcatWebServer  : Tomcat initialized with port(s): 8080 (http)
[indian-states] 2020-12-07 17:44:00.189  INFO 1 --- [           main] o.apache.catalina.core.StandardService   : Starting service [Tomcat]
[indian-states] 2020-12-07 17:44:00.196  INFO 1 --- [           main] org.apache.catalina.core.StandardEngine  : Starting Servlet engine: [Apache Tomcat/9.0.39]
[indian-states] 2020-12-07 17:44:00.387  INFO 1 --- [           main] o.a.c.c.C.[Tomcat].[localhost].[/]       : Initializing Spring embedded WebApplicationContext
[indian-states] 2020-12-07 17:44:00.388  INFO 1 --- [           main] w.s.c.ServletWebServerApplicationContext : Root WebApplicationContext: initialization completed in 5249 ms
[indian-states] 2020-12-07 17:44:01.673  INFO 1 --- [           main] com.zaxxer.hikari.HikariDataSource       : HikariPool-1 - Starting...
[indian-states] 2020-12-07 17:44:02.375  INFO 1 --- [           main] com.zaxxer.hikari.HikariDataSource       : HikariPool-1 - Start completed.
[indian-states] 2020-12-07 17:44:03.216  INFO 1 --- [           main] o.s.s.concurrent.ThreadPoolTaskExecutor  : Initializing ExecutorService 'applicationTaskExecutor'
[indian-states] 2020-12-07 17:44:04.050  INFO 1 --- [           main] o.s.b.w.embedded.tomcat.TomcatWebServer  : Tomcat started on port(s): 8080 (http) with context path ''
[indian-states] 2020-12-07 17:44:04.095  INFO 1 --- [           main] c.e.i.IndianStatesApplication            : Started IndianStatesApplication in 10.782 seconds (JVM running for 12.991)

As you can see that the application is built and deployed to the local Kubernetes cluster now. We first have to check the NodePort of our application using kubectl command to access the application locally.

kubectl get all
NAME                          READY   STATUS    RESTARTS   AGE
pod/states-7c55b8d5b6-vx5hq   1/1     Running   0          5m47s
NAME                 TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
service/kubernetes   ClusterIP   10.96.0.1        <none>        443/TCP          21h
service/states       NodePort    10.110.135.236   <none>        8080:30925/TCP   5m47s
NAME                     READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/states   1/1     1            1           5m48s
NAME                                DESIRED   CURRENT   READY   AGE
replicaset.apps/states-7c55b8d5b6   1         1         1       5m48s

NodePort assigned to our application is 30925. Let's invoke the /states REST endpoint of our application and see what happens.

curl localhost:30925/states
[{"name":"Andra Pradesh","capital":"Hyderabad"},{"name":"Arunachal Pradesh","capital":"Itangar"},{"name":"Assam","capital":"Dispur"},{"name":"Bihar","capital":"Patna"},{"name":"Chhattisgarh","capital":"Raipur"},{"name":"Goa","capital":"Panaji"},{"name":"Gujarat","capital":"Gandhinagar"},{"name":"Haryana","capital":"Chandigarh"},{"name":"Himachal Pradesh","capital":"Shimla"},{"name":"Jharkhand","capital":"Ranchi"},{"name":"Karnataka","capital":"Bangalore"},{"name":"Kerala","capital":"Thiruvananthapuram"},{"name":"Madhya Pradesh","capital":"Bhopal"},{"name":"Maharashtra","capital":"Mumbai"},{"name":"Manipur","capital":"Imphal"},{"name":"Meghalaya","capital":"Shillong"},{"name":"Mizoram","capital":"Aizawi"},{"name":"Nagaland","capital":"Kohima"},{"name":"Orissa","capital":"Bhubaneshwar"},{"name":"Rajasthan","capital":"Jaipur"},{"name":"Sikkim","capital":"Gangtok"},{"name":"Tamil Nadu","capital":"Chennai"},{"name":"Telangana","capital":"Hyderabad"},{"name":"Tripura","capital":"Agartala"},{"name":"Uttaranchal","capital":"Dehradun"},{"name":"Uttar Pradesh","capital":"Lucknow"},{"name":"West Bengal","capital":"Kolkata"},{"name":"Punjab","capital":"Chandigarh"}]

This looks great!!!

Let’s do a small code change and see if Skaffold can re-trigger the entire workflow. I will just change the replicas from 1 to 2 in a deployment YAML file and see if Skaffold can redeploy the application with increased replicas.

Application redeployed

Tags used in deployment:
 - docker.io/hiashish/indian-states -> docker.io/hiashish/indian-states:43f7c470a60b876c7579ed3041b64024b774e9808851ad83b6817701d0188cc5
Starting deploy...
 - deployment.apps/states configured
Waiting for deployments to stabilize...
 - deployment/states is ready.
Deployments stabilized in 4.109550864s
Watching for changes...
[indian-states] 
[indian-states]   .   ____          _            __ _ _
[indian-states]  /\\ / ___'_ __ _ _(_)_ __  __ _ \ \ \ \
[indian-states] ( ( )\___ | '_ | '_| | '_ \/ _` | \ \ \ \
[indian-states]  \\/  ___)| |_)| | | | | || (_| |  ) ) ) )
[indian-states]   '  |____| .__|_| |_|_| |_\__, | / / / /
[indian-states]  =========|_|==============|___/=/_/_/_/
[indian-states]  :: Spring Boot ::                (v2.4.0)
[indian-states]

Now check again replicaset for our application using kubectl.

kubectl get all
NAME                          READY   STATUS    RESTARTS   AGE
pod/states-7c55b8d5b6-br9zx   1/1     Running   0          20s
pod/states-7c55b8d5b6-vx5hq   1/1     Running   0          16m
NAME                 TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
service/kubernetes   ClusterIP   10.96.0.1        <none>        443/TCP          21h
service/states       NodePort    10.110.135.236   <none>        8080:30925/TCP   16m
NAME                     READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/states   2/2     2            2           16m
NAME                                DESIRED   CURRENT   READY   AGE
replicaset.apps/states-7c55b8d5b6   2         2         2       16m

As you can see after redeployment number of replicas has increased to 2 which was expected.

skaffold run

This is similar to skaffold dev but the main difference is that the workflow described in skaffold.yaml is executed just once. This is recommended for your production workflow.

Conclusion

In this article, I have just covered a handful of features that Skaffold provides. There are a lot of other features worth looking at like port-forwarding for debugging, health checks and templating deployment configurations, etc. It is worth considering Skaffold for continuous deployment of your Kubernetes applications.