Blueprint: OTC GitOps

This repository serves as a blueprint to start fast and easy with OTC GitOps

Requirements

First setup the infrastructure according to the README within this Github Template
then create a fork from this repository.

Introduction

This helm chart infrastructure-charts is automatically installed by terraform. It then creates multiple other applications in the format of app-of-apps pattern

Within infrastructure-charts/values.yaml you can add new services and customize them. Helm tpl is supported within the values.yaml file

Access ArgoCD UI

If you didn't source the shell-helper.sh in the https://github.com/iits-consulting/otc-terraform-template project please do so by running:

source shell-helper.sh

Now you are able to execute the argo command. Run the argo command. This will do the following:

1. Print out the Username and the Password on the first line
2. The browser should open automatically and open a tab to the ArgoCD UI. If it does not open a browser, you can do it yourself by opening this url: http://localhost:8080/argocd
3. You should see that ArgoCD automatically already installed multiple charts

If all services are up and running you should also be able to access your admin domain like this: https://admin.YOUR-DOMAIN-NAME

How to deploy some charts/services

You have 3 options to deploy some services.

1. Chart from a global helm chart registry which is configured in line number 12 (in this example we use https://charts.iits.tech/). Charts deployed like this: argocd-config, otc-storage-classes, traefik, cert-manager, basic-auth-gateway, kafka, admin-dashboard

2. Chart from a non global helm chart registry. Charts deployed like this: bitnami-kafka

3. Chart which resides inside this git repository. Charts deployed like this: akhq

Let's deploy some additional chart. Now it is time for you deploy some charts/services by yourself. In this example we will install an elastic stack (kibana/elasticsearch/filebeat)

• Open infrastructure-charts/values.yaml
• Add a new service like this:

  elastic-stack:
    namespace: monitoring
    targetRevision: "7.17.3-route-bugfix"

You need to commit and push this change now. Argo detects the changes and applies them after around 2-3 minutes.

After deployment please update the admin dashboard (infrastructure-charts/values-files/admin-dashboard/values.yaml) with the new links.

• /kibana
• /elasticsearch

If you don't want to search for icons you can see the solution here: https://github.com/iits-consulting/charts/blob/main/charts/iits-admin-dashboard/values.yaml

How to change values of the charts

You have 3 ways of changing the values of a chart

1. You change the values inside the remote/local helm chart itself
2. You set parameters inside the "infrastructure-charts/values.yaml" like shown between line number 55 till 57. We would recommend this approach if you need to template values or if you have just a few values which needs to be set.
3. You specify the location of a values.yaml file like shown on line number 82. We would recommend this approach only if you have a lot of static values which are not stage dependent.

Now let's change some values:

1. Please change inside /infrastructure-charts/values.yaml the number of traefik replicas from 1 to 2
2. Commit and push your changes
3. Check the service in the ArgoCD UI and verify that it scaled up

Handover variables from Terraform to ArgoCD

Since this setup is build on top of the otc-terraform-template you can hand over information from terraform to argo like this:

resource "helm_release" "argocd" {
  ...
  values                = [
    yamlencode({
      projects = {
        infrastructure-charts = {
          projectValues = {
            # Set this to enable stage $STAGE-values.yaml
            stage        = var.stage
            traefikElbId = module.terraform_secrets_from_encrypted_s3_bucket.secrets["elb_id"]
            adminDomain  = "admin.${var.domain_name}"
            storageClassKmsKeyId = module.terraform_secrets_from_encrypted_s3_bucket.secrets["storage_class_kms_key_id"]
          }
      ...
    }
    )
  ]
}

All projectValues variables are given over to argo, and we can reuse them here.

In this example the stage, traefikElbId, adminDomain storageClassKmsKeyId variables are handed over to argo.

How to integrate Business Apps

1. First copy the whole content of this project to some other git repository
2. Change then the folder infrastructure-charts to something you like for example app-charts
3. Change also all the other occurrences from infrastructure-charts to app-charts
4. Register the app-charts as a App of Apps project inside terraform like this:

resource "helm_release" "argocd" {
  ...
  values                = [
    yamlencode({
      projects = {
        infrastructure-charts = {
        ....
        }
        app-charts = {
          projectValues = {
            # Set this to enable stage $STAGE-values.yaml
            stage        = var.stage
            appDomain  = "${var.domain_name}"
          }

          git = {
            password = var.git_token
            repoUrl  = "https://my-git-repo-for-apps.git"
          }
        }
    }
    )
  ]
}

5. Argo will now do the same with the app-charts as with the infrastructure-charts

For each team we recommend to create a own git repo and AppProject. Then you will be able to fully make use of RBAC.

(Optional) Try out the setup

Now we go a little bit freestyle. Pick one of the topics below or choose one which you are interested in. Talk with your teammates and/or your tutor about it. Try to find the best way to implement it.

1. Setup a RDS database

   • How would you create a RDS?
   • How would you initialize the database with users,tables... ?
   • How can you avoid to work with IPs? Think about the thing that you need to set inside the microservice the private ip everytime.

2. Try to deploy a third party helm chart like bitnami-kafka

• Attention !!! only secured and hardened images are allowed to run inside our cluster the container could fail because of maybe root access is required
• Which steps need to be done to make thirdparty helm images secure?
• How would you develop a pipeline to sign you images and verify all images inside kubernetes are signed with a specific key?

3. Try to deploy your own nextcloud

   • Where do i store my data? What kind of persistent storage should i choose?
   • How do i backup my data?

4. Try to deploy a prometheus-stack

The End