KodeCloud ArgoCD

Links: 111 KodeCloud Index

Introduction - GitOps¶

GitOps process diagram:
In GitOps pipeline we have 2 repositories.
- One for the application code and other for the infrastructure.
Difference between GitOps and DevOps:

In GitOps changes were pulled by the operator and in DevOps changes were pushed to the clutser.

Difference between push and pull based approach:

The main drawback of a push based approach is that cluster config is embedded in the CI system.

A single GitOps operator can deploy applications to multiple clusters without the need for it to be installed or setup on all the clusters.
Some Benefits of GitOps:
- It is vendor neutral.
- Git is the source of truth.
- Helps in eliminating configuration drift.
- We have revisions of the infrastructure with git history.

ArgoCD¶

Argo CD is a declarative, GitOps continuous delivery tool for Kubernetes resources defined in a Git repository.
- Continuously monitors running applications and comparing their live state to the desired state
- It reports the deviations and provides visualizations to help developers manually or automatically sync the live state with the desired state.
Terminologies:
- Application: It is a custom resource definition which represents a deployed application in k8s.
  - It has 2 main components
    - Source repository URL
    - Destination cluster
  - Application source type: The application can have different source types. E.g. Helm, Kustomize or Ksonnet
  - Diagram:
- Project: Provide a logical grouping of applications, which is useful when Argo CD is used by multiple teams.
  - By default a default project is created for us in ArgoCD which is very permissive.
- Target state: The desired state of an application, as represented by files in a Git repository.
- Live state: The live state of that application.
  - What pods, configmap, secrets, etc are created/deployed in a Kubernetes cluster.
- Sync status: Whether or not the live state matches the target state.
  - Is the deployed application the same as Git says it should be?
  - In case of rollbacks the sync status will always be out of sync.
- Sync: The process of making an application move to its target state.
  - E.g. by applying changes to a Kubernetes cluster.
- Refresh: Compare the latest code in Git with the live state.
  - Figure out what is different.
- Health: The health of the application, is it running correctly? Can it serve requests?
  - We can create custom health checks.
    - For example this is particularly useful since config maps have no predefined health check.
  - Diagram:
ArgoCD Architecture:

Both the UI and the CLI talks to the ArgoCD Rest API.

ArgoCD installation types:
If we want to use ArgoCD cli for accessing its API we have to install it separately.
ArgoCD has a reconciliation loop which by default runs every 3 minutes to poll the data from the source git server.
- This can be changed by adding information in the argocd configmap and then restarting the ArgoCD deployment.
Diagram for understanding the difference between sync and refresh:
We can also use a webhook based approach if we want to remove the polling delay.

Sync Strategies¶

When ArgoCD detects a new version of our application in git it either performs a manual sync or an automated sync.
- If set to automatic, Argo CD will apply the changes then update or create new resources in the target Kubernetes cluster.
- If set to manual, Argo CD WON'T make any changes until and unless a user clicks the sync button on the UI.
Auto-pruning feature describes what happens when files are deleted or removed from Git.
- If this feature disabled then argocd WON'T delete anything from the cluster.
Enabling the self heal feature will ensure that any manual edits made to the cluster will be reverted back to the state described in Argo CD.
By default syncing is set to manual and auto pruning and self healing is disabled.

Best configuration: manual sync with auto pruning and self healing enabled.

Can I only enable auto pruning and self healing in argocd without enabling automatic sync?¶

To enable auto pruning, you can set the --auto-prune flag to true when starting the Argo CD server: argocd-server --auto-prune=true
To enable self-healing, you can set the --self-heal flag to true when starting the Argo CD server: argocd-server --self-heal=true
We can enable auto pruning and self-healing in Argo CD without enabling automatic sync by setting the respective flags as mentioned above while leaving the sync behaviour unchanged.

Declarative Setup¶

We can create ArgoCD applications using UI or ArgoCD cli.
We can also use k8s like manifests to create applications ArgoCD.
App of apps pattern:
App of apps manages other ArgoCD applications.

Multi Cluster Deployment¶

Argo CD can be used to manage deployments in multiple cluster.
Argo CD gets the cluster credentials from the kubeconfig file.
So we need to update out kubeconfig file to have the details of the external server.
Once we have the cluster credentials in the kubeconfig file we need to add the cluster to argocd
- argocd cluster add <context-name>
We can list the clusters using: argocd cluster list
Credentials to external clusters are stored as secrets within the Argo CD namespace.
Diagram:

Miscellaneous¶

User management:

Last updated: 2023-06-07