Set endpoint as environment variable in Azure Container Instance

Setting the endpoint of a Container Instance as an environment variable is possible, but the endpoint is not immediately available at the time of container creation. Below are two workarounds you can use to achieve this

Workaround 1

you can set environment variables when creating the container instance but construct the actual endpoint URL within your application after the container has started.

When you create your container using CLI, PowerShell, or Portal, you can define environment variables that might include placeholders for the endpoint.

az container create \
  --resource-group arkorg \
  --name arkocontainer \
  --image mcr.microsoft.com/oss/nginx/nginx:1.21.6 \
  --environment-variables ENDPOINT_PLACEHOLDER="http://<placeholder>" \
  --dns-name-label arkocontainer \
  --ports 80

or construct the endpoint in your application Logic i.e. after the container starts, your app could fetch the container’s environment variables and then modify the placeholder to the actual URL.

As discussed in comments, to retrieve the full FQDN of the container after it's been deployed. In your case, the unique identifier (hbh3g4ly23g4yh3h) is part of the FQDN and can be extracted using:

az container show --resource-group arkorg --name arkocontainer --query ipAddress.fqdn --output tsv

Then construct the full URL in your Python Webhook

example-

import os
import subprocess

# Step 1: Retrieve the full FQDN using Azure CLI
fqdn = subprocess.check_output(
    ["az", "container", "show", "--resource-group", "arkorg", "--name", "arkocontainer", "--query", "ipAddress.fqdn", "--output", "tsv"],
    universal_newlines=True
).strip()

# Step 2: Construct the URL
if fqdn:
    protocol = "http"  # Or "https" if using SSL
    full_endpoint = f"{protocol}://{fqdn}"

    # Now use this endpoint for your webhook
    print(f"Webhook Endpoint: {full_endpoint}")
    
    # Assuming this is the URL you need to send out
    send_webhook_to(f"{full_endpoint}/webhook-reply")  # Example function
else:
    print("FQDN could not be retrieved.")

Workaround 2

Dynamically set the endpoint as an environment variable post deployment i.e. first, deploy your container as usual, without setting the endpoint environment variable. After deployment, you can retrieve the container’s endpoint using-

az container show --resource-group arkorg --name arkocontainer --query ipAddress.fqdn --output tsv

Now you can use some sort of startup script within your container to query the endpoint and set it as an environment variable.

Example-

# startup.sh
ENDPOINT=$(az container show --resource-group arkorg --name arkocontainer --query ipAddress.fqdn --output tsv)
export CONTAINER_ENDPOINT=$ENDPOINT

# Run your application
exec "$@"

and in your Dockerfile, set this script to run at startup

ENTRYPOINT ["/path/to/startup.sh"]

Personally, I would suggest Workaround 1 as it is more straightforward if your application can handle constructing the endpoint after startup and can work with placeholders.

MS docs-

• Set environment variables in container instances
• Set environment variables in container instances