Attaching a load balancer to a GKE cluster
Load balancers are a must-have for any containerized application that wants to run on a cluster. You can easily add a load balancer and specify the pods to which it should direct traffic. This makes it easy to repeatedly deploy new services while always making them accessible via the load balancer, thus reducing down time.
Shippable does not directly integrate GKE load balancers (services) in
deploy jobs. The main reason for this is that GKE's labeling system makes it unnecessary to couple the two pieces. Instead, you can connect a GKE service to your replicationController at any time by referencing the Shippable pod labels, or by adding your own custom labels to your manifest.
1. Add a loadBalancer resource.
resources: - name: app_lb type: loadBalancer integration: gke pointer: sourceName: "apploadbalancer" method: LoadBalancer namespace: shippable clusterName: "multiple-zones-cluster" region: "us-east1-b" version: ports: - name: testport protocol: TCP port: 80 selector: name: "app_service_def" jobName: "app_deploy_job"
The loadBalancer resource supports a range of fields, allowing you to set almost any field that you'd want. Complete reference is here.
Shippable adds the values of
jobName as labels to every pod spec for GKE. These need to be set as follows:
name: "name of your manifest job that was deployed"
jobName: "name of the deploy job that deployed your manifest"
By using both the manifest name and job name as selectors, the GKE service will always pinpoint the exact pods from the replicationController that was deployed.
If you want to set custom labels for your loadBalancer, check out the instructions to specify custom labels.
2. Add a
The provision job is used to create ancillary objects like load balancers on Container Orchestration Platforms like GKE and Kubernetes.
Add the following yml block to your shippable.yml file.
jobs: - name: app_provision #friendly name for job type: provision steps: - IN: app_lb #name of load balancer you created in step 1
This job will use the
integration associated with your
loadBalancer resource to make API calls to your cluster to create the loadBalancer on your GKE cluster with all of the settings you requested.
You can use a single provision job to provision multiple loadBalancers. Just include a separate
IN statement with a separate resource for each service you plan to create.
Setting custom labels
If you have another set of labels that you'd like to use instead, you can use a dockerOptions resource to set your own labels. Labels will be set on the pods.
resources: - name: app_labels type: dockerOptions version: labels: name: "api" environment: "test"
app_labelsas an additional input to the manifest job. For example, the yml snippet for our single container application would look like this:
jobs: - name: app_service_def type: manifest steps: - IN: app_image - IN: app_options - IN: app_environment - IN: app_labels
- Update the
selectorsection in the
loadBalancerresource to match the labels you used in
dockerOptions. The yml snippet for our single container application would look like this:
resources: - name: app_lb type: loadBalancer integration: op_int pointer: sourceName: "apploadbalancer" method: LoadBalancer namespace: shippable clusterName: "multiple-zones-cluster" region: "us-east1-b" version: ports: - name: testport protocol: TCP port: 80 selector: #this should match the labels in dockerOptions name: "api" environment: "test"
Ask questions on Chat
Feel free to engage us on Chat if you have any questions about this document. Simply click on the Chat icon on the bottom right corner of this page and someone from our customer success team will get in touch with you.
Improve this page
We really appreciate your help in improving our documentation. If you find any problems with this page, please do not hesitate to reach out at firstname.lastname@example.org or open a support issue. You can also send us a pull request to the docs repository.