A guide to deploy with Spinnaker

A Continuous Integration and Continuous Deployment (CI/CD) pipeline is an automated process that helps deploy applications with little to zero manual effort. CI/CD pipelines are a critical part of a software development lifecycle, as a well-thought-out pipeline can foster speedy software delivery.

While there are several tools out there, both commercial and open-source, to help set up your CI/CD pipeline, in this article, we look at a particular tool called Spinnaker.

Developed by Netflix as an open-source, mult-icloud continuous delivery tool, Spinnaker was released as a successor to Netflix’s internally developed tool Asgard in November 2015. It supports most cloud-native platforms, including AWS, Azure, GCP, Kubernetes, and Oracle Cloud; Google currently also helps extend and maintain it.

Features of Spinnaker

Key features of Spinnaker include:

  • Frequent updates and maintenance from Netflix, in collaboration with Google
  • Integration with all major cloud providers; support for seamless multi-cloud continuous delivery
  • Ability to configure a wide variety of automated triggers; support for numerous notification delivery integrations, including Slack and Gmail
  • Ability to scale, proven by the fact that major organizations around the world, including SAP and Adobe, have adopted it

Setting up Spinnaker on Kubernetes

We will be installing Spinnaker on minikube using the Docker driver. Minikube is a single-node minimal Kubernetes cluster widely used for proof of concept purposes. You can set up a minikube cluster by following this guide.

Since Spinnaker consumes significant resources, i.e., a minimum of 4 cores and 8 GB RAM, let’s ensure that our minikube instance has that allocated:

minikube start --memory 8192 --cpus 4

Once minikube is up and running, you can install Halyard as a Docker container. Halyard is specially designed for configuring and managing Spinnaker. You need to install it in the minikube network because it will be interacting with the minikube instance:

docker run --name halyard -v ~/.hal:/home/spinnaker/.hal -v 
~/.kube/config:/home/spinnaker/.kube/config --network="minikube" -
d gcr.io/spinnaker-marketplace/halyard:stable

Get a shell inside the Halyard container and confirm that you can run kubectl commands. Once we have Kubernetes connectivity, enable the Kubernetes cloud provider:

hal config provider kubernetes enable

For Spinnaker to be able to access objects in the cluster, you will need a Kubernetes account, which we have enabled with the account name “spinnaker-test” using the following command:

hal config provider kubernetes account add spinnaker-test --
provider-version v2 --context $(kubectl config current-context)

Now, associate your Kubernetes account with Halyard:

hal config deploy edit --type distributed --account-name 

Next, configure a persistent storage for Spinnaker. Here we’re using S3 buckets to enable persistence but you can use MinIO as well:

hal config storage s3 edit \ 
--access-key-id "<Access Key ID>" \
--secret-access-key \
--region us-west-2
hal config storage edit –type s3

You will also need to choose a version of Spinnaker to install:

hal config version –list 
hal config version edit –version <selected-version>

Finally, to install Spinnaker:

hal deploy apply

Getting all the Spinnaker pods up and running will take some time. Once they’re up, we need to make some final configuration changes with regard to the dashboard. To access the dashboard from your browser, convert the spin-deck and spin-gate services in the Spinnaker namespace as nodeport services, where spin-deck service is the actual UI endpoint of spinnaker and spin-gate service is the gateway endpoint which needs to interact with the UI endpoint and set the URLs in the Halyard config.

To expose the nodeport services:

kubectl patch svc spin-deck --type='json' -p 
'[{"op":"replace","path":"/spec/type","value":"NodePort"},{"op":"replace","path":"/spec/ports/0/nodePort","value":30010}]' -n
kubectl patch svc spin-gate --type='json' -p
'[{"op":"replace","path":"/spec/type","value":"NodePort"},{"op":"replace","path":"/spec/ports/0/nodePort","value":30020}]' -n

To set the URLs:

hal config security ui –edit –override-base-url 
hal config security api –edit –override-base-url
hal deploy apply

Deploying a pipeline with Spinnaker

To create a basic CD pipeline, where we can deploy NGINX and a clusterIP service leveraging Spinnaker, navigate to the dashboard and click on “Create Application”:

Fill in the details for the service you want to deploy. For our example, we only need to fill in the first two blocks:

Click on “Create”.

Now, your application will show up on the Applications page. Click on it, and then select “Configure a new pipeline”:

Now, we can add a stage and fill in the details, including the account name and type of stage, which in this case would be “Deploy (Manifest)”. You can also add the deployment file as shown below:

The deployment’s config will be:

apiVersion: v1 
kind: Service
name: nginx-svc
namespace: default
app: nginx
- port: 80
targetPort: 80
apiVersion: apps/v1
kind: Deployment
name: nginx-deployment
namespace: default
replicas: 2
app: nginx
app: nginx
- image: nginx:1.14.2
name: nginx-container
containerPort: 80

Save the configuration, and navigate to the Pipelines page. Manually trigger the pipeline, and watch the bar as the pipeline completes in the background:

Your NGINX deployment pipeline is now complete and the NGINX app should be deployed in the default namespace.

While Spinnaker is largely used for deploying applications that have pre-built artifacts, it can also be integrated with CI tools such as Jenkins to add build stages in the pipeline. This way, you can plan an entire pipeline around Spinnaker and take advantage of its features.

The Good?

Compared to other pipeline-based tooling, Spinnaker has certain key advantages:

  • It is exceptionally well documented and maintained by two organizations—Netflix and Google—known for developing good tooling and for reviewing/considering any bugs or requests for new features, provided it benefits their tool.
  • It offers seamless integration with all known cloud providers and other CI tools.
  • It is feature-rich compared to other pipeline tools currently available; plus, all the latest features and add-ons are free (no paid commercial version).
  • It has easy-to-use support for automated and instant rollbacks in case of failure.

The Bad?

While Spinnaker is an extremely useful and powerful tool, there are certain things you might want to keep an eye on:

  • It is extremely resource-heavy, even to get it started; thus, it might not be suitable if you want to minimize on-demand costs.
  • It can be a little complicated to install and get started with, especially in a Windows environment; this tool is unsuitable for beginners.
  • It is insecure by default, meaning you need to spend some time securing it.


In this blog, we’ve installed Spinnaker from scratch by leveraging Halyard to deploy a sample NGINX deployment in minikube. As seen above, you can easily integrate Spinnaker with major cloud providers, not just Kubernetes.

One of the best use cases for Spinnaker is canary deployments for effective and segregated pipelines to keep your deployments updated with the latest changes.

Considering the pros and cons of Spinnaker is critical for engineers to determine if it’s suitable for the organization and will help them architect effective CI/CD pipelines.

Was this article helpful?

Write For Us

Write for Site24x7 is a special writing program that supports writers who create content for Site24x7 "Learn" portal. Get paid for your writing.

Write For Us

Write for Site24x7 is a special writing program that supports writers who create content for Site24x7 “Learn” portal. Get paid for your writing.

Apply Now
Write For Us