Imported from jrojas-pi/pi-acr-manager (SKILL.md). Install upstream with npx skills add jrojas-pi/pi-acr-manager. Copyright stays with the author.

PI ACR Manager

End-to-end workflow for building Docker images, pushing them to Azure Container Registry, and updating Azure Container Apps.

Workflow

Step 1 — Gather parameters

Before doing anything, collect these values. Use vscode_askQuestions to ask the user for all of them at once:

For version detection, search for __version__.py in the workspace using file_search with pattern **/__version__.py. Parse the file and extract the value assigned to __version__. This is the image tag.

Step 2 — Set Azure subscription

az account set --subscription "<subscription>"

Step 3 — Resolve ACR login server

az acr show --name <acr_name> --resource-group <resource_group> --query loginServer -o tsv

Save the result as LOGIN_SERVER — you'll need it for tagging and pushing.

Step 4 — Build and push the image

Try local Docker first, then fall back to cloud build.

Option A — Local Docker build

# Login to ACR
az acr login --name <acr_name>

# Build
docker build -t <image_name>:<version> .

# Tag
docker tag <image_name>:<version> <LOGIN_SERVER>/<image_name>:<version>

# Push
docker push <LOGIN_SERVER>/<image_name>:<version>

If az acr login or docker build fail because Docker is not running or not installed, switch to Option B.

Option B — Cloud build (ACR Tasks)

This builds the image directly in the cloud — no local Docker needed.

Before running, check if a .dockerignore file exists. If not, create one to exclude heavy/unnecessary directories:

.venv
.git
.vscode
.agents
__pycache__
*.pyc
.env
.env-*
env.local
scripts
*.md
*.docx
.gitignore
.funcignore
.dockerignore
docker-compose.yml
azure-pipelines.yml

Then check the Dockerfile — if it pulls from Docker Hub (e.g., FROM python:3.12-slim), Docker Hub rate limits may block the cloud build. To handle this:

1. Check if the base image already exists in the ACR:

   az acr repository list --name <acr_name> -o table
   az acr repository show-tags --name <acr_name> --repository <base_image_name> -o table

2. If the base image exists in ACR, temporarily update the Dockerfile FROM line to reference the ACR copy:

   FROM <LOGIN_SERVER>/<base_image_name>:<tag>
   

3. Run the cloud build:

   az acr build --registry <acr_name> --resource-group <resource_group> --image <image_name>:<version> --file Dockerfile .

4. Restore the Dockerfile to its original FROM line after the build completes (whether it succeeded or failed). The Dockerfile in the repo must always reference the upstream image, not the ACR-specific one.

If the base image doesn't exist in ACR and Docker Hub rate limits hit, try importing it first:

az acr import --name <acr_name> --source docker.io/library/<image>:<tag> --image <image>:<tag> --force

Step 5 — Verify image in ACR

az acr repository show-tags --name <acr_name> --repository <image_name> -o table

Confirm the version tag appears in the list.

Step 6 — Update the Container App

az containerapp update \
  --name <container_app> \
  --resource-group <resource_group> \
  --image <LOGIN_SERVER>/<image_name>:<version>

Step 7 — Verify deployment

Run these three checks:

1. Image in use:

   az containerapp show --name <container_app> --resource-group <resource_group> --query "properties.template.containers[0].image" -o tsv

   Confirm it matches <LOGIN_SERVER>/<image_name>:<version>.

2. Active revision:

   az containerapp revision list --name <container_app> --resource-group <resource_group> --query "[-1].{name:name, active:properties.active, replicas:properties.replicas, trafficWeight:properties.trafficWeight, createdTime:properties.createdTime}" -o table

   Confirm Active=True and TrafficWeight=100.

3. Image exists in ACR:

   az acr repository show-tags --name <acr_name> --repository <image_name> -o table

Step 8 — Final summary

Present a summary to the user in this format:

## Resumen de despliegue

- **Imagen**: <image_name>:<version>
- **ACR**: <acr_name> (<LOGIN_SERVER>)
- **Digest**: <sha256 digest from build output>
- **Container App**: <container_app>
- **Revisión activa**: <revision name>
- **Réplicas**: <count>
- **Tráfico**: <weight>%
- **Estado**: ✅ Desplegado exitosamente / ❌ Error (con detalle)

Error handling

• If any az command fails, show the error and suggest potential fixes (wrong subscription, permissions, resource not found)
• If Docker Hub rate limit is hit, use the ACR-cached base image strategy described in Step 4 Option B
• If the Container App update fails, check if the ACR credentials are configured in the Container App
• Always restore the Dockerfile if it was temporarily modified