Using Scoped Variables in Devtron

Introduction

In any piece of software or code, variables are used for holding data such as numbers or strings. Variables are created by declaring them, which involves specifying the variable's name and type, followed by assigning it a value.

Devtron offers super-admins the capability to define scoped variables (key-value pairs). It means, while the key remains the same, its value may change depending on the following context:

Global: Variable value will be universally same throughout Devtron.
Cluster: Variable value might differ for each Kubernetes cluster.
Environment: Variable value might differ for each environment within a cluster, e.g., staging, dev, prod.
Application: Variable value might differ for each application.
Environment + Application: Variable value might differ for each application on a specific environment.

Advantages of using scoped variables

Reduces repeatability: Configuration management team can centrally maintain the static data.
Simplifies bulk edits: All the places that use a variable get updated when you change the value of the variable.
Keeps data secure: You can decide the exposure of a variable's value to prevent misuse or leakage of sensitive data.

How to Define a Scoped Variable

On Devtron, a super-admin can download a YAML template. It will contain a schema for defining the variables.

Download the Template

From the left sidebar, go to Global Configurations → Scoped Variables
Click Download template.
Open the downloaded template using any code editor (say VS Code).

Enter the Values

The YAML file contains key-value pairs that follow the below schema:

Field	Type	Description
`apiVersion`	string	The API version of the resource (comes pre-filled)
`kind`	string	The kind of resource (i.e. Variable, comes pre-filled)
`spec`	object	The complete specification object containing all the variables
`spec.name`	string	Unique name of the variable, e.g. DB_URL
`spec.shortDescription`	string	A short description of the variable (up to 120 characters)
`spec.notes`	string	Additional details about the variable (will not be shown on UI)
`spec.isSensitive`	boolean	Whether the variable value is confidential (will not be shown on UI if true)
`spec.values`	array	The complete values object containing all the variable values as per context

The spec.values array further contains the following elements:

Field	Type	Description
`category`	string	The context, e.g., Global, Cluster, Application, Env, ApplicationEnv
`value`	string	The value of the variable
`selectors`	object	A set of selectors that restrict the scope of the variable
`selectors.attributeSelectors`	object	A map of attribute selectors to values
`selectors.attributeSelectors.<selector_key>`	string	The key of the attribute selector, e.g., ApplicationName, EnvName, ClusterName
`selectors.attributeSelectors.<selector_value>`	string	The value of the attribute selector

Here's a truncated template containing the specification of two variables for your understanding:

apiVersion: devtron.ai/v1beta1
kind: Variable
spec:

# First example of a variable
  - name: DB_URL
    shortDescription: My application's customers are stored
    notes: The DB is a MySQL DB running version 7.0. The DB contains confidential
      information.
    isSensitive: true
    values:
      - category: Global
        value: mysql.example.com

# Second example of a variable
  - name: DB_Name
    shortDescription: My database name to recognize the DB
    notes: NA
    isSensitive: false
    values:
      - category: Global
        value: Devtron
      - category: ApplicationEnv
        value: app1-p
        selectors:
          attributeSelectors:
            ApplicationName: MyFirstApplication
            EnvName: prod

Upload the Template

Once you save the YAML file, go back to the screen where you downloaded the template.
Use the file uploader utility to upload your YAML file.
The content of the file will be uploaded for you to review and edit. Click Review Changes.
You may check the changes between the last saved file and the current one before clicking Save.
Click the Variable List tab to view the variables. Check the How to Use a Scoped Variable section to know more.

How to Edit an Existing Scoped Variable

Only a super-admin can edit existing scoped variables.

Option 1: Directly edit using the UI

Option 2: Reupload the updated YAML file

{% hint style="warning" %} Reuploading the YAML file will replace the previous file, so any variable that existed in the previous file but not in the latest one will be lost {% endhint %}

How to Use a Scoped Variable

Once a variable is defined, it can be used by your authorized users on Devtron. A scoped variable widget would appear only on the screens that support its usage.

Currently, the widget is shown only on the following screens in App Configuration:

Workflow Editor → Edit build pipeline → Pre-build stage (tab)
Workflow Editor → Edit build pipeline → Post-build stage (tab)
Workflow Editor → Edit deployment pipeline → Post-Deployment stage (tab)
Workflow Editor → Edit deployment pipeline → Post-Deployment stage (tab)
Deployment Template
ConfigMaps
Secrets

Upon clicking on the widget, a list of variables will be visible.

Use the copy button to copy a relevant variable of your choice.

It would appear in the following format upon pasting it within an input field: @{{variable-name}}

Order of Precedence

When multiple values are associated with a scoped variable, the precedence order is as follows, with the highest priority at the top:

Environment + App
App
Environment
Cluster
Global

Example

Environment + App: This is the most specific scope, and it will take precedence over all other scopes. For example, the value of DB name variable for the app1 application in the prod environment would be app1-p, even though there is a global DB name variable set to Devtron. If a variable value for this scope is not defined, the App scope will be checked.
App: This is the next most specific scope, and it will take precedence over the Environment, Cluster, and Global scopes. For example, the value of DB name variable for the app1 application would be project-tahiti, even though the value of DB name exists in lower scopes. If a variable value for this scope is not defined, the Environment scope will be checked.
Environment: This is the next most specific scope, and it will take precedence over the Cluster and Global scopes. For example, the value of DB name variable in the prod environment would be devtron-prod, even though the value of DB name exists in lower scopes. If a variable value for this scope is not defined, the Cluster scope will be checked.
Cluster: This is the next most specific scope, and it will take precedence over the Global scope. For example, the value of DB name variable in the gcp-gke cluster would be Devtron-gcp, even though there is a global DB name variable set to Devtron-gcp. If a variable value for this scope is not defined, the Global scope will be checked.
Global: This is the least specific scope, and it will only be used if no variable values are found in other higher scopes. The value of DB name variable would be Devtron.

List of Predefined Variables

There are some system variables that exist by default in Devtron that you can readily use if needed:

DEVTRON_NAMESPACE: Provides name of the namespace
DEVTRON_CLUSTER_NAME: Provides name of the cluster configured on Devtron
DEVTRON_ENV_NAME: Provides name of the environment
DEVTRON_IMAGE_TAG: Provides image tag associated with the container image
DEVTRON_IMAGE: Provides full image path of the container image, e.g., gcr.io/k8s-minikube/kicbase:v0.0.39
DEVTRON_APP_NAME: Provides name of the application on Devtron

{% hint style="info" %} Currently, these variables do not appear in the scoped variable widget, but you may use them. {% endhint %}

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Using Scoped Variables in Devtron

Introduction