Skip to content

Deploy to Amazon ECS #2225

Deploy to Amazon ECS

Deploy to Amazon ECS #2225

Workflow file for this run

# This workflow will build and push a new container image to Amazon ECR,
# and then will deploy a new task definition for each of the services to Amazon ECS.
# Expected vars:
# The following vars must be set at the repo level. Their values must be JSON and contain one key per
# environment (dev, prod, etc.) and the value for each key must be the value for that environment.
#
# DEPLOY_APP_NAME: {"dev": "app1", "prod": "app2"}
# DEPLOY_AWS_REGION: {"dev": "us-west-2", "prod": "us-east-1"}
# AWS_ACCOUNT: {"dev": "123456789012", "prod": "123456789012"}
# Note: The names of repository, cluster, services match what is configured in https://github.com/dimagi/ocs-deploy
name: Deploy to Amazon ECS
on:
workflow_dispatch:
inputs:
environment:
description: "Deploy environment"
required: true
type: choice
options:
- dev
- prod
workflow_run:
workflows: [ Lint and Test ]
types: [completed]
branches: [main]
permissions:
id-token: write
contents: read
deployments: write
concurrency:
group: ${{ github.workflow }}-${{ github.ref }}-${{ inputs.environment }}
cancel-in-progress: false
jobs:
deploy:
name: Deploy
runs-on: ubuntu-latest
if: ${{ github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success' }}
steps:
- name: Checkout
uses: actions/checkout@v6
- name: Set Deploy Env
# Set the deploy env based on the input from the event or else from the branch
run: |
if [[ "${{ github.event_name }}" == "workflow_dispatch" ]]; then
DEPLOY_ENV="${{ inputs.environment }}"
elif [[ "${{github.base_ref}}" == "main" || "${{github.ref}}" == "refs/heads/main" ]]; then
DEPLOY_ENV="prod"
else
DEPLOY_ENV="dev"
fi
echo "DEPLOY_ENV=$DEPLOY_ENV" >> "$GITHUB_ENV"
- name: Set variables
# Set other variables accordingly
run: |
# you can't reference the `env` context when defining other env vars to do it here
APP_NAME="${{ fromJSON(vars.DEPLOY_APP_NAME)[env.DEPLOY_ENV] }}"
AWS_REGION="${{ fromJSON(vars.DEPLOY_AWS_REGION)[env.DEPLOY_ENV] }}"
{
echo "APP_NAME=$APP_NAME"
echo "AWS_REGION=$AWS_REGION"
echo "ECR_REPOSITORY=$APP_NAME-${{ env.DEPLOY_ENV }}-ecr-repo"
echo "ECS_CLUSTER=$APP_NAME-${{ env.DEPLOY_ENV }}-Cluster"
echo "ECS_SERVICE_DJANGO=$APP_NAME-${{ env.DEPLOY_ENV }}-Django"
echo "ECS_SERVICE_CELERY=$APP_NAME-${{ env.DEPLOY_ENV }}-Celery"
echo "ECS_SERVICE_CELERY_BEAT=$APP_NAME-${{ env.DEPLOY_ENV }}-CeleryBeat"
echo "DJANGO_STACK_NAME=$APP_NAME-${{ env.DEPLOY_ENV }}-$AWS_REGION-django-stack"
} >> "$GITHUB_ENV"
- name: Create GitHub deployment
uses: chrnorm/deployment-action@v2
id: deployment
with:
token: '${{ github.token }}'
environment: "aws-${{ env.DEPLOY_ENV }}"
production-environment: ${{ env.DEPLOY_ENV == 'prod' }}
description: "Deploying ${{ github.head_ref || github.ref }} to AWS ${{ env.DEPLOY_ENV }}"
- name: configure aws credentials
uses: aws-actions/configure-aws-credentials@v6.2.1
with:
role-to-assume: "arn:aws:iam::${{ fromJSON(vars.AWS_ACCOUNT)[env.DEPLOY_ENV] }}:role/github_deploy"
role-session-name: GithubDeploy
aws-region: ${{ fromJSON(vars.DEPLOY_AWS_REGION)[env.DEPLOY_ENV] }}
- name: Login to Amazon ECR
id: login-ecr
uses: aws-actions/amazon-ecr-login@v2.1.6
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Get image names
id: image-name
env:
ECR_REGISTRY: ${{ steps.login-ecr.outputs.registry }}
IMAGE_TAG: ${{ github.sha }}
run: |
echo "image=$ECR_REGISTRY/$ECR_REPOSITORY:$IMAGE_TAG" >> "$GITHUB_OUTPUT"
echo "image_latest=$ECR_REGISTRY/$ECR_REPOSITORY:latest" >> "$GITHUB_OUTPUT"
- name: Check for pre-built image
# The Lint and Test workflow pre-builds the prod image (tagged by SHA) in
# parallel with the tests, so the common main->prod deploy can skip the
# build entirely. Manual dev deploys (and the rare case where the pre-build
# hasn't finished) fall through to the build step below.
id: check-image
run: |
if aws ecr describe-images \
--repository-name "$ECR_REPOSITORY" \
--image-ids imageTag="${{ github.sha }}" >/dev/null 2>&1; then
echo "exists=true" >> "$GITHUB_OUTPUT"
echo "Image ${{ github.sha }} already in ECR — skipping build"
else
echo "exists=false" >> "$GITHUB_OUTPUT"
echo "Image ${{ github.sha }} not found in ECR — building"
fi
- name: Build and push
if: steps.check-image.outputs.exists != 'true'
uses: docker/build-push-action@v7
with:
tags: |
${{ steps.image-name.outputs.image }}
${{ steps.image-name.outputs.image_latest }}
file: Dockerfile
cache-from: type=gha
cache-to: type=gha,mode=max
outputs: type=image,push=true,oci-mediatypes=true,compression=zstd,compression-level=3
- name: Update ECS task def for Django web container
id: django-web-def
uses: aws-actions/amazon-ecs-render-task-definition@v1.8.4
with:
task-definition-family: ${{ env.APP_NAME }}-${{ env.DEPLOY_ENV }}-Django
container-name: web
image: ${{ steps.image-name.outputs.image }}
- name: Update ECS task def for Migration container
id: migration-def
uses: aws-actions/amazon-ecs-render-task-definition@v1.8.4
with:
task-definition-family: ${{ env.APP_NAME }}-${{ env.DEPLOY_ENV }}-Migration
container-name: migrate
image: ${{ steps.image-name.outputs.image }}
- name: Update ECS task def for Celery worker container
id: celery-worker-def
uses: aws-actions/amazon-ecs-render-task-definition@v1.8.4
with:
task-definition-family: ${{ env.APP_NAME }}-${{ env.DEPLOY_ENV }}-CeleryWorkerTask
container-name: celery-worker
image: ${{ steps.image-name.outputs.image }}
- name: Update ECS task def for Celery beat container
id: celery-beat-def
uses: aws-actions/amazon-ecs-render-task-definition@v1.8.4
with:
task-definition-family: ${{ env.APP_NAME }}-${{ env.DEPLOY_ENV }}-CeleryBeatTask
container-name: celery-beat
image: ${{ steps.image-name.outputs.image }}
- name: Get stack outputs for network config
id: stack-outputs
run: |
set -e
OUTPUTS=$(aws cloudformation describe-stacks \
--stack-name "$DJANGO_STACK_NAME" \
--query 'Stacks[0].Outputs' \
--output json) || { echo "Failed to retrieve CloudFormation stack outputs"; exit 1; }
SUBNETS=$(echo "$OUTPUTS" | jq -r '.[] | select(.OutputKey | endswith("PrivateSubnets")) | .OutputValue')
SG=$(echo "$OUTPUTS" | jq -r '.[] | select(.OutputKey | endswith("ServiceSecurityGroup")) | .OutputValue')
[[ -z "$SUBNETS" ]] && { echo "Error: Could not find subnets in stack outputs"; exit 1; }
[[ -z "$SG" ]] && { echo "Error: Could not find security group in stack outputs"; exit 1; }
echo "subnets=$SUBNETS" >> "$GITHUB_OUTPUT"
echo "security_group=$SG" >> "$GITHUB_OUTPUT"
- name: Run migration task
id: run-migration
run: |
# Strip metadata fields that cause issues when re-registering
jq 'del(.tags, .registeredAt, .registeredBy, .taskDefinitionArn, .revision, .status, .requiresAttributes, .compatibilities)' \
${{ steps.migration-def.outputs.task-definition }} > /tmp/migration-task-def.json
# Register task definition with new image
TASK_DEF_ARN=$(aws ecs register-task-definition \
--cli-input-json file:///tmp/migration-task-def.json \
--query 'taskDefinition.taskDefinitionArn' \
--output text)
# Run the migration task
TASK_ARN=$(aws ecs run-task \
--cluster "$ECS_CLUSTER" \
--task-definition "$TASK_DEF_ARN" \
--launch-type FARGATE \
--network-configuration "awsvpcConfiguration={subnets=[${{ steps.stack-outputs.outputs.subnets }}],securityGroups=[${{ steps.stack-outputs.outputs.security_group }}],assignPublicIp=DISABLED}" \
--query 'tasks[0].taskArn' \
--output text)
echo "task_arn=$TASK_ARN" >> "$GITHUB_OUTPUT"
echo "Started migration task: $TASK_ARN"
- name: Wait for migration & output logs
run: |
echo "Waiting for migration task to complete..."
aws ecs wait tasks-stopped \
--cluster "$ECS_CLUSTER" \
--tasks "${{ steps.run-migration.outputs.task_arn }}"
# Check exit code
EXIT_CODE=$(aws ecs describe-tasks \
--cluster "$ECS_CLUSTER" \
--tasks "${{ steps.run-migration.outputs.task_arn }}" \
--query 'tasks[0].containers[0].exitCode' \
--output text)
# Get task details for log configuration
TASK_ID=$(echo "${{ steps.run-migration.outputs.task_arn }}" | cut -d'/' -f3)
# Output migration logs
LOG_STREAM="${LOG_STREAM_PREFIX}/${TASK_ID}"
echo "Fetching migration logs from CloudWatch..."
LOGS=$(aws logs get-log-events \
--log-group-name "$LOG_GROUP" \
--log-stream-name "$LOG_STREAM" \
--limit 100 \
--query 'events[*].message' \
--output json 2>&1) || LOGS="Could not fetch logs"
# Output to console
echo "----------------------------------------"
echo "$LOGS"
echo "----------------------------------------"
# Add to job summary
{
echo "## Migration Logs"
echo ""
echo "**Task ID:** \`$TASK_ID\`"
echo "**Log Group:** \`$LOG_GROUP\`"
echo "**Log Stream:** \`$LOG_STREAM\`"
echo ""
echo "<details>"
echo "<summary>View Logs</summary>"
echo ""
echo "\`\`\`json"
echo "$LOGS"
echo "\`\`\`"
echo "</details>"
} >> "$GITHUB_STEP_SUMMARY"
if [ "$EXIT_CODE" != "0" ]; then
echo "Migration failed with exit code: $EXIT_CODE"
{
echo ""
echo "❌ **Migration failed with exit code: $EXIT_CODE**"
} >> "$GITHUB_STEP_SUMMARY"
exit 1
fi
echo "Migration completed successfully"
{
echo ""
echo "✅ **Migration completed successfully**"
} >> "$GITHUB_STEP_SUMMARY"
env:
LOG_GROUP: ${{ env.APP_NAME }}-${{ env.DEPLOY_ENV }}-DjangoMigrationLogs
LOG_STREAM_PREFIX: ${{ env.APP_NAME }}-${{ env.DEPLOY_ENV }}-migrate/migrate
- name: Deploy Django Web
uses: aws-actions/amazon-ecs-deploy-task-definition@v2
with:
task-definition: ${{ steps.django-web-def.outputs.task-definition }}
service: ${{ env.ECS_SERVICE_DJANGO }}
cluster: ${{ env.ECS_CLUSTER }}
wait-for-service-stability: false
- name: Deploy Celery Worker
uses: aws-actions/amazon-ecs-deploy-task-definition@v2
with:
task-definition: ${{ steps.celery-worker-def.outputs.task-definition }}
service: ${{ env.ECS_SERVICE_CELERY }}
cluster: ${{ env.ECS_CLUSTER }}
wait-for-service-stability: false
- name: Deploy Celery Beat
uses: aws-actions/amazon-ecs-deploy-task-definition@v2
with:
task-definition: ${{ steps.celery-beat-def.outputs.task-definition }}
service: ${{ env.ECS_SERVICE_CELERY_BEAT }}
cluster: ${{ env.ECS_CLUSTER }}
wait-for-service-stability: false
- name: Wait for service stability
run: |
# Run the three waiters in parallel. `set -e` (default in GH Actions bash)
# doesn't apply to background processes, so we capture each exit code via
# `wait <pid>` and fail only after all three have finished.
aws ecs wait services-stable --cluster "$ECS_CLUSTER" --services "$ECS_SERVICE_DJANGO" &
pid_django=$!
aws ecs wait services-stable --cluster "$ECS_CLUSTER" --services "$ECS_SERVICE_CELERY" &
pid_celery=$!
aws ecs wait services-stable --cluster "$ECS_CLUSTER" --services "$ECS_SERVICE_CELERY_BEAT" &
pid_celery_beat=$!
rc=0
wait "$pid_django" || { echo "Django service failed to stabilise"; rc=1; }
wait "$pid_celery" || { echo "Celery service failed to stabilise"; rc=1; }
wait "$pid_celery_beat" || { echo "CeleryBeat service failed to stabilise"; rc=1; }
exit $rc
- name: Update deployment status
if: always()
uses: chrnorm/deployment-status@v2
with:
token: '${{ github.token }}'
environment-url: ${{ steps.deployment.outputs.environment_url }}
deployment-id: ${{ steps.deployment.outputs.deployment_id }}
state: ${{ job.status == 'success' && 'success' || 'failure' }}
- name: Create Sentry release
if: success()
uses: getsentry/action-release@v3
env:
SENTRY_AUTH_TOKEN: ${{ secrets.SENTRY_AUTH_TOKEN }}
SENTRY_ORG: ${{ secrets.SENTRY_ORG }}
SENTRY_PROJECT: ${{ secrets.SENTRY_PROJECT }}
with:
environment: "${{ env.DEPLOY_ENV == 'prod' && 'production' || 'development' }}"
ignore_empty: true
ignore_missing: true