Deploying to Kubernetes Permalink to "Deploying to Kubernetes"

JHipster app(s) and associated services can be deployed in the following ways

  • Standard kubectl/kustomize/skaffold sub-generator jhipster kubernetes|k8s
  • Helm sub-generator jhipster kubernetes-helm|helm-k8s
  • Knative sub-generator jhipster kubernetes-knative|knative

jhipster kubernetes | k8s Permalink to "jhipster kubernetes | k8s"

This sub-generator generates manifests that is deployed to Kubernetes through kubectl/kustomize/skaffold cli.

Limitations Permalink to "Limitations"

  • Cassandra is not supported yet
  • Kubernetes v1.9+ is required

Pre-requisites Permalink to "Pre-requisites"

You have to install:

You must have a Docker registry. If you don’t have one, you can use the official Docker Hub.

Minikube Permalink to "Minikube"

Minikube is a tool that helps to run Kubernetes locally. Minikube runs a single-node Kubernetes cluster inside a VM on your laptop for users looking to try out Kubernetes or develop with it day-to-day.

You can use it to test your application before pushing it to Kubernetes.

Running the sub-generator Permalink to "Running the sub-generator"

To generate config files for Kubernetes, run this command in a new folder:

jhipster kubernetes | k8s

Then answer all the questions to deploy your application.

Which type of application would you like to deploy? Permalink to "Which type of application would you like to deploy?"

Your type of application depends on whether you wish to deploy a microservices architecture or classical applications.

Enter the root directory where your applications are located Permalink to "Enter the root directory where your applications are located"

Enter the path.

Which applications do you want to include in your Kubernetes configuration? Permalink to "Which applications do you want to include in your Kubernetes configuration?"

Select your applications.

Do you want to setup monitoring for your applications? Permalink to "Do you want to setup monitoring for your applications?"

Select the option.

Enter the admin password used to secure the JHipster Registry admin Permalink to "Enter the admin password used to secure the JHipster Registry admin"

This question is only displayed if you choose microservices architecture with JHipster Registry.

What should we use for the Kubernetes namespace? Permalink to "What should we use for the Kubernetes namespace?"

See the documentation on namespace here

What should we use for the base Docker repository name? Permalink to "What should we use for the base Docker repository name?"

If you choose Docker Hub as main registry, it will be your Docker Hub login.

If you choose Google Container Registry, then it’ll be[PROJECT ID], or a regional registry, such as[PROJECT ID],[PROJECT ID], or[PROJECT ID]. See Pushing and Pulling Images for more detial.

If you choose other registries like Harbor, Quay or something similar, then the login would be like <registry_server>/<repo>/[PROJECT ID]

What command should we use for push Docker image to repository? Permalink to "What command should we use for push Docker image to repository?"

The default command to push to Docker Hub is docker image push

If you use Google Container Registry to publish your Docker images, it will be: gcloud docker push

Choose the Kubernetes service type for your edge services? Permalink to "Choose the Kubernetes service type for your edge services?"

Select the appropriate K8s routing type.

These are the standard prompts. In addition, other prompts will be shown based on choosing options like Istio, Ingress etc.

Updating your deployed application Permalink to "Updating your deployed application"

Preparing a new deployment Permalink to "Preparing a new deployment"

When your application is already deployed, you can re-deploy it by building a new Docker image:

./mvnw package -Pprod -DskipTests jib:dockerBuild

Or when using Gradle:

./gradlew -Pprod bootJar jibDockerBuild -x test

Pushing to Docker Hub Permalink to "Pushing to Docker Hub"

Tag locally your image:

docker image tag application username/application

Push your image to Docker Hub:

docker image push username/application

Deploying a monolith/microservice application Permalink to "Deploying a monolith/microservice application"

You can deploy all your apps by running the below command:

./ -f (default option)  [or] ./ -k (kustomize option) [or] ./ -s (skaffold run)

You can deploy apps using kustomize:

kubectl apply -k ./

You can deploy apps using skaffold:

skaffold run [or] skaffold deploy

It will deploy your apps and its associated dependent services (database, elasticsearch, etc.).

Custom namespaces Permalink to "Custom namespaces"

It is possible to specify a custom namespace for the entire deployment. To perform custom commands, you have to specify the target namespace, like in this example:

kubectl get pods -n <custom-namespace>

Scaling your deployments Permalink to "Scaling your deployments"

You can scale your applications using

kubectl scale deployment <app-name> --replicas <replica-count>

Zero downtime deployments Permalink to "Zero downtime deployments"

The default way to update a running application in Kubernetes is to deploy a new image tag to your Docker registry and then deploy it using:

kubectl set image deployment/<app-name>-app <app-name>=<new-image>

Using livenessProbes and readinessProbe allows you to tell Kubernetes about the state of your applications, in order to ensure availability of your services. You will need a minimum of 2 replicas for every application if you want to have zero downtime deployment. This is because the rolling upgrade strategy first stops a running replica in order to place a new one. Running only one replica will cause a short downtime during upgrades.

Deploying a Service Registry in Kubernetes Permalink to "Deploying a Service Registry in Kubernetes"

Although Kubernetes features its own internal service discovery with Kube-DNS, JHipster rely on Spring Cloud for service discovery, so it depends on a third party service registry like Eureka or Consul. This has the advantage of being platform independent and to work similarly in production and on a local development machine.

Consequently, for microservices applications, the JHipster Kubernetes sub-generator will generate Kubernetes manifest files to deploy service registries like the JHipster-Registry (based on Eureka) or Consul. Moreover, the generated microservices and gateway Kubernetes manifests will contains the appropriate configuration to register themselves to their central registry.

Managing the JHipster Registry or Consul in Kubernetes Permalink to "Managing the JHipster Registry or Consul in Kubernetes"

For the JHipster Registry and Consul, StatefulSets configurations are provided. Those are a certain kind of Kubernetes resources which can handle stateful applications and will let you scale your service registries for high availability. For more information on high-availability for Eureka and Consul refer to their respective documentation.

Centralized configuration in Kubernetes Permalink to "Centralized configuration in Kubernetes"

Centralized configuration is also setup using either Spring Cloud Config Server (when using the JHipster Registry) or the Consul Key/Value store (when using Consul). By default, both configuration servers load their configuration from a Kubernetes ConfigMap which contains property files in this format:

apiVersion: v1
kind: ConfigMap
  name: application-config
  namespace: default
  application.yml: |- # global properties shared by all applications
            secret: secret
  gateway-prod.yml: |- # gateway application properties for the "prod" profile
      bar: foobar

By default, configuration servers run in development mode, which means that YAML property files are read directly from the filesystem and hot-reloaded on changes. For production it is advised to setup configuration from a Git repository as explained in our microservice documentation for the JHipster-Registry config server and Consul config server.

Exposing headless services Permalink to "Exposing headless services"

The registry is deployed using a headless service in Kubernetes, so the primary service has no IP address, and cannot get a node port. You can create a secondary service for any type, using:

kubectl expose service jhipster-registry --type=NodePort --name=exposed-registry

and explore the details using

kubectl get svc exposed-registry

For scaling the JHipster Registry, use

kubectl scale statefulset jhipster-registry --replicas 3

Monitoring tools Permalink to "Monitoring tools"

The sub-generator provides monitoring tools and configuration for usage with your applications.

Prometheus metrics Permalink to "Prometheus metrics"

If not already done, install the Prometheus operator by CoreOS. You can quickly deploy the operator using

kubectl create -f

hint: More information on how to enable and protect prometheus metrics in your application you can find in our monitoring documentation.

The Prometheus instance for your applications can be explored using

kubectl get svc prometheus-appname

Taking Advantage of Kubernetes Permalink to "Taking Advantage of Kubernetes"

Kubernetes offers a number of facilities out-of-the-box to help with Microservices deployments, such as:

  • Service Registry - Kubernetes Service is a first-class citizen that provides service registry and lookup via DNS name.
  • Load Balancing - Kubernetes Service acts as a L4 load balancer.
  • Health Check - Liveness probes and readiness probes help determine the health of the service.
  • Configuration - Kubernetes ConfigMap can be used to store and apply configuration outside of the application.

There are a number of benefits of using Kubernetes facilities:

  • Simplified deployment
  • No need for additional Eureka/Consul deployment
  • No need for Spring Cloud Gateway to proxy/route requests
  • No need for Spring Cloud Load Balancer

At the same time, there are some drawbacks:

  • No Application Management through JHipster Registry - This function relies on Spring Cloud’s DiscoveryClient. This can be updated in the future to add spring-cloud-kubernetes
  • No local Docker Compose support - You must use minikube for local development, and use Ingress to route traffic
  • No request-level load balancing - Kubernetes Service is a L4 load balancer that load balances per connection. Use Istio for request level load balancing (see below).

Using Kubernetes as Service Registry Permalink to "Using Kubernetes as Service Registry"

To avoid relying on Eureka or Consul, you’ll need to disable service discovery altogether

  • When asked Which service discovery server do you want to use?, choose No service discovery

A JHipster Gateway usually fronts the API calls and routing these calls using Spring Cloud Gateway. Without a service registry, routing via Spring Cloud Gateway won’t work. You’ll need to use Kubernetes Ingress to route the traffic to microservices.

  • When asked Choose the kubernetes service type for your edge services, choose Ingress.

Istio Permalink to "Istio"

You can deploy microservices into Istio-enabled Kubernetes cluster. While Kubernetes manages microservices deployment and configuration, Istio can manage service to service communication, such as request-level load balancing, retries, circuit breakers, traffic routing/splitting, and more.

To enable Istio support:

  • When asked Do you want to configure Istio?, choose one of the Istio options
  • When asked Do you want to generate Istio route files, choose Yes to generate default configuration for circuit breaking, etc.

Troubleshooting Permalink to "Troubleshooting"

My applications don’t get pulled, because of ‘imagePullBackoff’

Check the registry your Kubernetes cluster is accessing. If you are using a private registry, you should add it to your namespace by kubectl create secret docker-registry (check the docs for more information).

My applications get stopped, before they can boot up

This can occur, if your cluster has low resources (e.g. Minikube). Increase the initialDelySeconds value of livenessProbe of your deployments.

My applications are starting very slow, despite I have a cluster with many resources

The default setting is optimized for middle-scale clusters. You are free to increase the JAVA_OPTS environment variable, resource requests and limits to improve the performance. Be careful!

I have selected Prometheus but no targets are visible

This depends on the setup of Prometheus operator and the access control policies in your cluster. Version 1.6.0+ is needed for the RBAC setup to work.

I have selected Prometheus, but my targets never get scraped

This means your applications are probably not built using the prometheus profile in Maven/Gradle

My SQL-based microservice are stuck during Liquibase initialization when running multiple replicas

Sometimes the database changelog lock gets corrupted. You will need to connect to the database using kubectl exec -it and removes all lines of liquibases databasechangeloglock table.

jhipster kubernetes-helm | k8s-helm Permalink to "jhipster kubernetes-helm | k8s-helm"

This sub-generator generates manifests that is deployed to Kubernetes through helm cli.

Prerequisites Permalink to "Prerequisites"

To use the manifest generated by this sub-generator helm cli should be installed. Follow this link for installation instructions. This requires helm 2.12.x or later.

Once Helm is installed you need to add the below repositories:

helm repo add stable
helm repo add incubator

These repositories should be added to the local cache, because this sub-generator will pull stable production grade services charts from the above repos.

This sub-generator uses the kubernetes sub-generator for the app(s) manifests and pulls services like database, elasticsearch, prometheus etc. images being referred in your app(s) from the above repositories.

Deployment Permalink to "Deployment"

You can deploy all your app(s) by running the following command:

bash (or) ./ will always do a clean install. Any of the existing charts with the same identity gets removed first and then it does a clean install.

You can upgrade all your apps (if you have made any changes to the generated manifests) by running the following command:

bash (or) ./

jhipster kubernetes-knative | knative Permalink to "jhipster kubernetes-knative | knative"

This sub-generator generates manifests that is deployed to Kubernetes through either kubectl or helm cli. It generates manifests for one of the cli’s based on the chosen prompt response.

Prerequisites Permalink to "Prerequisites"

This sub-generator depends on Istio. In order to use the manifests generated by this sub-generator, you should have istio and kntaive installed in the cluster. Follow this link for installation instructions. It requires knative 0.8.x or later.

Deployment Permalink to "Deployment"

If you have chosen to deploy using Kubernetes generator, run the below command:

bash (or) ./

If you have chosen to deploy using Helm generator, run the below command:

bash (or) ./ will always do a clean install. Any of the existing charts with the same identity gets removed first and then it does a clean install.

You can upgrade all your apps (if you have made any changes to the generated manifests) by running the following bash command:

bash (or) ./

More information Permalink to "More information"