Skip to content

Kubernetes.md

Latest commit

 

History

History
99 lines (87 loc) · 3.59 KB

Kubernetes.md

File metadata and controls

99 lines (87 loc) · 3.59 KB

Container orchestration provides us:

  • Auto Scaling: Scale containers based on demand
  • Service Discovery: help microservices to find each other
  • Load Balancer: Distribute load among multiple instances of a microservice
  • Self Healing: Do health checks and replace failing instances
  • Zero Downtime Deployments: Release new versions without downtime

We install gcloud and kubectl to the local machine.
Then remove some dependencies in microservices. Because eureka, config, api server etc. provided by kubernetes. Also add some config for liveness and readiness probes.

Create deployment

kubectl create deployment currency-exchange --image=guzelcihad/mmv2-currency-exchange-service:0.0.11-SANPSHOT

Expose service

kubectl expose deployment currency-exchange --type=LoadBalancer --port=8000

PS

Remember we use feign client to make api calls like this

@FeignClient(name="currency-exchange")

It was ok with the docker setup. Now in Kubernetes we expose services to communicate so we need to provide hostname.
If you have a service called currency-exchange then Kubernetes expose an environment variable called CURRENCY_EXCHANGE_SERVICE_HOST and assign this variable to PODs. We use this variable to specify hostname in Feign client like this.

@FeignClient(name = "currency-exchange", url = "${CURRENCY_EXCHANGE_SERVICE_HOST:http://localhost}:8000")

This simply say that, if I find CURRENCY_EXCHANGE_SERVICE_HOST than I am gonna make a request this url, otherwise I am gonna make request to localhost:8080.

Go into the project folder and type to create deployment and service files

kubectl get deployment currency-exchange -o yaml >> deployment.yaml
kubectl get service currency-exchange -o yaml >> service.yaml

Then copy all the info from service.yaml and add "---" to the end of the deployment file and paste service yaml info. This way we merge all the objects into one file.
Increase replica 1 to 2 in deployment file.
To see the difference type

kubectl diff -f deployment.yaml

And update

kubectl apply -f deployment.yaml

Enable Logging and Tracing API's in Google Cloud

  • Search APIs and Services
  • Enable api and services
  • Enable cloud logging api
  • Enable all stack drivers api

Communication Between Microservices

Instead of using kubernetes supplied environment variables, we can create our own variables. First, change the name of the env variable in feign client class.
Then open deployment.yaml file and create env. variables.

We define it under containers info.

    spec:
      containers:
      - image: in28min/mmv2-currency-conversion-service:0.0.12-SNAPSHOT
        imagePullPolicy: IfNotPresent
        name: mmv2-currency-conversion-service
        env:
          - name: CURRENCY_EXCHANGE_URI
            value: http://currency-exchange

Host name refers to service name. This way kubernetes can distribute the load.
We just harcoded define env. variables instead of that we use configmap.

kubectl create configmap currency-conversion --from-literal=CURRENCY_EXCHANGE_URI=http://currency-exchange

Liveness and Readiness Probes

  • If readiness probe is not successful, no traffic is sent
  • If liveness proble is not successful, pod is restarted

Spring boot actuator (>= 2.3) provides inbuilt readiness and liveness probes

  • health/readiness
  • health/liveness

We add the actuator dependency to the project and add the specified configs in application.properties

management.endpoint.health.probes.enabled=true
management.health.livenessState.enabled=true
management.health.readinessState.enabled=true