container-apps-dapr

Example project for working with Container Apps with DAPR and Terraform

This repository contains a sample project for working with Container Apps with DAPR and Terraform.

• Getting started
  • Prerequisites
• Provision environment
• Build
• Deploy
  • Push images to Azure Container Registry
  • Deploy Container Apps
• Usage
  • Example with curl

Getting started

Prerequisites

• Azure account (with at least Contributor role on a resource group)
• Go >=1.21
• Terraform >=1.4.0
• Azure CLI >=2.45.0

Download the project:

git clone https://github.com/RedeployAB/container-apps-dapr.git

Provision environment

Create a variables file for the environment like so:

touch deployments/terraform/application-environment/terraform.tfvars

Add the following variables to the file:

// If creating a new resource group, set to true.
deploy_resource_group = false

resource_group_name = "<name-of-resource-group>"
location            = "<location>"
identity_name       = "<name-of-identity>"

// Set this to create a new workspace.
log_analytics_workspace_name  = "<name-of-log-analytics>"
 // Set this to deploy to an existing workspace.
log_analytics_workspace_id    = "<log-analytics-workspace-id>"

storage_account_name    = "<name-of-storage-account>"
servicebus_name         = "<name-of-servicebus-namespace>"
container_registry_name = "<name-of-azure-container-registry>"

container_app_environment_name = "<name-of-container-app-environment>"
provision_container_apps       = false

// Messaging system to use. Options are "queue" or "pubsub".
messaging_system      = "queue"
// Name of the applications in the project.
messaging_dapr_scopes = ["endpoint", "worker"]
// Name of the "worker" application of the project.
output_dapr_scopes    = ["worker"]

Run Terraform:

cd deployments/terraform/application-environment
terraform plan -out=tfplan

# Verify the output and apply.
terraform apply tfplan

Build

# Build the endpoint binary.
./scripts/bash/build.sh --module endpoint --version 1.0.0 --image

# Build the worker binary.
./scripts/bash/build.sh --module worker --version 1.0.0 --image

Deploy

Push images to Azure Container Registry

# Login to the container registry.
az acr login --name <name-of-azure-container-registry>

# Tag the endpoint and push it to the registry.
docker tag endpoint:1.0.0 <name-of-azure-container-registry>.azurecr.io/endpoint:1.0.0
docker push <name-of-azure-container-registry>.azurecr.io/endpoint:1.0.0

# Tag the worker and push it to the registry.
docker tag worker:1.0.0 <name-of-azure-container-registry>.azurecr.io/worker:1.0.0
docker push <name-of-azure-container-registry>.azurecr.io/worker:1.0.0

Deploy Container Apps

Terraform

Edit deployments/terraform/application-environment/terraform.tfvars and add:

provision_container_apps = true
endpoint_security_keys   = ["<key>"]

Run Terraform:

cd deployments/terraform/application-environment
terraform plan -out=tfplan

# Verify the output and apply.
terraform apply tfplan

Script

Note: Use the Terraform way of provisioning instead, this script was created when there where certain limitiations with the Terraform modules and container app scaling.

uuid=$(uuidgen)

./deployment/scripts/deploy-container-apps.sh \
  --resource-group <name-of-resource-group> \
  --environment <name-of-container-app-environment> \
  --identity <name-of-identity> \
  --registry <name-of-azure-container-registry>.azurecr.io \
  --servicebus-namespace <name-of-servicebus-namespace> \
  --messaging-system queue \  # Supports 'pubsub' deployments. Defaults to 'queue' if omitted.
  --endpoint-version 1.0.0 \
  --worker-version 1.0.0 \
  --endpoint-api-keys $uuid

Usage

POST /reports

{
  "id": "12345",
  "data": "testdata" // base64 encoded
}

Example with curl

# Get the URL for the `endpoint` container:
url=https://$(az containerapp show \
  --resource-group <resource-group-name> \
  --name endpoint \
  --query 'properties.configuration.ingress.fqdn' \
  --output tsv
)

# The $uuid should contain either a key set in the variable endpoint_security_keys,
# or the same $uuid as was used with the script deployment.
data=$(echo 'testdata' | base64)
curl -H "X-API-Key: $uuid" $url/reports --data "{\"id\":\"12345\",\"data\":\"$data\"}"