This walkthrough will cover every step needed to get your docker image built and deployed through FlexDeploy. In this example we will be deploying the image to a Kubernetes Cluster on Oracle Cloud Infrastructure via Helm.
If you don’t have your own docker application you can follow along using the Natours repository used in this example.
Docker CLI - An available docker cli for building the image
Kubernetes Cluster - For deployment
This guide will be using OKE, but any cluster will suffice.
Be sure the appropriate context is configured for your K8s Cluster.
We won’t be covering some of the more basic steps like creating SCM Instances and other Topology configuration.
The first goal is to build our image. In the case of the Natours repository, we don’t need any additional compile steps. All we need to do is run docker build with the appropriate commands
The build workflow is only going to need two steps: clone the source code and build the image. The buildProjectImage step is going to use the configuration directly in the project. As such, the default configuration is is fine for both steps, simply drag and drop.
If we are going to be pushing our image anywhere, we will need a Registry Account in FlexDeploy. Navigate to Topology->Integrations->Containers and click the Plus button.
Be sure to select the Provider as DockerRegistry and populate the information as necessary. In the above account it is configured to push to the joelwenzel account on dockerhub.com
While still on the Topology screen, we can create the remaining items for the build process.
Create an endpoint where Docker is installed
If docker is installed on the FlexDeploy server, the localhost endpoint can be used.
Create a Docker instance
Add the Build Environment to the Instance
Ensure that the Docker Instance has the workflow created above added. This is how FlexDeploy knows what properties to show on certain screens.
Docker on Windows or Docker-Desktop
If you are using docker on windows, you will need to perform these additional steps.
Add the Docker Endpoint to your Docker-Build EnvInst in Topology Overview
Create the SCM Instance for the Git repository
Lastly, create a new Project in FlexDeploy. Give it a name that closely resembles your Git repository name and select Container as the Classification.
Update the Project Configuration tab to include the Build Workflow, Build Instance and SCM Configuration.
After creating the Project, click on the Container Configuration tab. It is here where we specify the image tag information and build behavior.
Add the following configuration:
Naturally your image name and account could be different compared to what’s shown above. With this configuration, FlexDeploy will build an image with an incrementing Project Version every time we run a build. In addition to tagging the image with the ProjectVersion it will tag with latest and push the image to the Docker Hub registry that was set up earlier.
We will come back to image scanning, but for now lets finally build the image.
Click the Build button on the Project Activity screen and if everything goes well we should have a new version of the natours image.
Notice that both version 1.0.1 and the latest tag were both pushed to the registry.
A reminder that before deployment, you should have kubectl and helm installed on the local FlexDeploy server and have a valid context configured. Similar to the build process, we will first configure the project.
The only step our deploy workflow needs to do is run Helm, which we will do through the deploy operation on the FlexDeployHelmPlugin. The only input specified is the K8s Context which should be the kubectl context as it shows when running
kubectl config current-context
More of the same as what we did on the Build side of things.
Create a Helm Instance
Associate the DeployHelm workflow
Set the Localhost endpoint on the Development Environment for the Helm Instance.
Associate the Deploy Workflow and Deploy Instance on the Project Configuration tab.
Almost there, the final piece of the puzzle is to update the Helm configuration on the project.
Release Name - "natours-"+FD_ENVIRONMENT_CODE.toLowerCase()
This chart will create a single pod deployment with a load balancer service.
These will likely change depending on your chart and docker image.
Head back to the Project Activity Screen and click the deploy button. If everything goes well there will be a new successful deployment of the Helm Release.
You may notice a few things that stand out compared to other FlexDeploy Projects, namely the Deployment link and Status. Both of these are shown for Container Projects only and offer some insight into the actual application that was deployed and some basic monitoring.
The deployment status is only monitored if Kubectl is installed on the local FlexDeploy server.
If you have made it this far you have seen the bread and butter of Kubernetes deployments in FlexDeploy. If you want to take things to the next level this section will incorporate Image scanning via Anchore in our build process.
FlexDeploy supports native Anchore inline image scanning. What this means for you is that no additional installations or setup is needed apart from updating the Project configuration.
At this time, inline(local) image scanning is not supported on windows
Navigate back to the Container Configuration tab on the Project. Under the Image Scanning section, change the Scan Image drop down to Before Push and set the Fail When script to
STATUS == “fail”.
We have just told FlexDeploy to run an Anchore scan prior to pushing our image and if the scan Status is “fail” then the build will fail. Save the changes and run another build. If you are using the natours repository, the build will now fail due to the scan results.