Azure Container Instances for Single Container Deployments

Introduction

Deploying Containers has been a widespread trend in many businesses in recent years. Applications are Containerized and deployed on the Cloud. Containerization provides an operating system and application isolation. These can even get scaled automatically based on the traffic it gets. There are many cloud platforms that developers can work with to host running Containers. These Containers can be anything from APIs to data science apps to static websites. GCP, AWS, and Microsoft Azure are the well-known cloud platforms that provide these services.  In this article, we will look at one of Azure’s services called Azure Container Instances that developers can work with to deploy Containers in the cloud.

Learning Objectives

• Understand the workings of Azure Container Instances
• Learn how to push Images to Azure Container Registry
• Learn how to create Resource Groups
• Learn how to deploy Containers to Azure Container Instances

This article was published as a part of the Data Science Blogathon.

Table of Contents

1. What are Azure Container Instances?
2. Azure Container Instance – Pricing
3. Creating a Streamlit Application
4. Creating Azure Container Registry and Pushing the Docker Image
5. Hosting the Streamlit Application in Azure Container Instances
6. Conclusion

What are Azure Container Instances?

Microsoft Azure’s Azure Container Instances (ACI) cloud service allows users to run containers in the cloud without maintaining virtual machines. It provides a quick and seamless way for deploying and managing Containers on Azure. It is important to make it perfect for various situations, including simple apps, automating tasks, and building jobs. When compared to virtual machines, Containers are faster during launch. Without the need to provision and manage virtual machines, ACI may launch Containers in Azure in a matter of seconds. This makes it possible for containerized apps to be deployed quickly and effectively.

With an IP address and a Fully Qualified Domain Name (FQDN), users of ACI can immediately expose their container groups to the internet. Additionally, users can give a unique DNS name label to make their applications visible at a unique URL. ACI has an interactive shell that allows running commands in a container and can be used for application development and troubleshooting. Container access is protected using HTTPS over TLS for communication with Clients.

Azure Container Instance: Pricing

The Azure Container Instance follows the pay-for-compute-capacity by the second rule. We are billed for every computational second the Container is run. And the prices for these computations purely rely on the number of vCPUs we have allocated and the memory allocated to them. There is no upfront cost involved. It even depends on the operating system that the Container is built on, i.e. Linux or Windows. Windows Containers tend to have higher prices when compared to the Linux Containers.

And finally, note that the prices for computations defer from Region to Region. For example, if the location selected is East US and the operating system is Linux, the computation costs would be something like the below:

Memory Cost: $0.0000013 Per GB-s

CPU Cost: $0.0000113 Per vCPU-s

These costs displayed are per-second costs. These costs are then multiplied by the total seconds the Container has run and the final bill is produced. And note that the bill generated will be lower because these Containers only run when they receive a Request, thus low costs.

Creating a Streamlit Application

In this section, we will be creating a Streamlit application in Python. After creating the application, we will convert it into an Image and host it in the Container Instance. Before starting this, make sure you have an active Microsoft Azure account. Do not worry, you will not be charged, as when creating an Azure account, we get a free $200 to use.

The streamlit app we are going to create is a simple application where the user inputs a range of values for the X variable. Then the user selects a function to calculate Y from X. Like if the selects the function Y = log(X), then the Y will contain the log values for each element present in X. Like these, we will provide 4 different functions, which include log(), sin(), e^(), and tan().

The code for creating the streamlit application described above can be referred to below:

import streamlit as st
import pandas as pd
import numpy as np

# Setting up the website title and icon
st.set_page_config(page_title="Data Visualization",
                   page_icon=":chart_with_upwards_trend:")

# setting the title name
st.title('Multi Function Graph')

# user inputs the x values and they are converted into a list
st.write("### X Data")
x = st.text_input(label="Provide X values with spaces").split()
x = list(map(int,x))

# user selects the function to calculate y
func = st.selectbox("Function: ",['y = log(x)', 'y = sin(x)', 'y = e^(x)', 'y = tan(x)'])

# the function is then run to get the y values
if func == 'y = log(x)':
    changed_y = [np.log2(x), 'y = log(x)']

elif func == 'y = sin(x)':
    changed_y = [np.sin(x), 'y = sin(x)']

elif func == 'y = e^(x)':
    changed_y = [np.exp(x), 'y = e^(x)']

elif func == 'y = tan(x)':
    changed_y = [np.tan(x), 'y = tan(x)']

# dataframe is constructed on X and Y data and the plot is made
data = pd.DataFrame({'x':x,changed_y[1]:changed_y[0]})
st.write("### X vs Y Graph")
st.line_chart(data=data,x='x',y=changed_y[1])

Let’s test our streamlit application by running it with the below command:

$ streamlit run app.py

Now we will be able to access the website application that we have created:

Now let’s provide values for x and keep the function y = log(x) and see the output:

Now for the same x values let’s keep the function y = tan(x) and see the below graph:

Let’s write a docker file to containerize our streamlit application. The Dockerfile will look like the below:

# Setting the Base Image
FROM python:3.7.9-slim

# Copy all files from current directory to app directory
COPY . /app

# Change working directory to app directory
WORKDIR /app

# Run below command to install all the libraries
RUN pip install -r requirements.txt

# Exposes PORT 8080
EXPOSE 80

#Running the application
ENTRYPOINT ["streamlit", "run", "--server.port=80","--server.enableCORS","false", "app.py"]

The following Dockerfile will be used to create the Image for the application. The base Image being used is the python:3.7.9-slim. Then we install the dependencies present in the requirements.txt using the pip. Here we expose PORT 80 because we want our streamlit application to run at PORT 80. Finally, the ENTRYPOINT runs the command provided for the application to run. The requirements.txt will contain the following:

• streamlit
• pandas

Creating Azure Container Registry and Pushing the Docker Image

In this section, we will create a Resource Group. After that, we will create an Azure Container Registry to push our streamlit Image to it. To create the Resource Group, click on the Resource Group Icon on the Dashboard(), click on Create, and then provide the resource name. After providing the resource name, click on the Next button present below and finally click on the Create button.

After creating the Resource Group, now we will create a Container Registry to store our Images. The Container Registry can be found through the Search Box. Click on it to create a new Container Registry. We have named the ACR streamregistry (Registry Name). Select the Resource Group that we have created before. Follow the Images and keep clicking on Next (let the Configuration be set to default), until we reach the Create button. Then finally click on Create button to create our Container Registry.

After clicking on Create, the deployment process for creating the Azure Container Registry will start. After it’s completed, the screen will show something similar to the below pic. Click on Go to Resource. Finally, we have our Container Registry ready.

Now, it’s time to build the Image. So use the below commands in the command line to build the Image for the streamlit application:

$ docker build -t streamregistry.azurecr.io/stream_function_graph .
$ az login
$ az acr login -n streamregistry 
$ docker push streamregistry.azurecr.io/stream_function_graph

• In the first command, replace streamlitregistry with the name of your Azure Container Registry. And replace stream_function_graph with the name of your application. This will then build our Image (Don’t forget the dot(.) at the end of the command).
• The second command will prompt you to log in to your Azure account.
• Then, through the third command, we log into the Container Registry that we have just created.
• Finally, the fourth command will push the Image to the Azure Container Registry

Now, in the above section, we have created the Azure Container Registry. In that, in the left pane, we will find a service called Repositories. Click on that; we will be able to view the Image that we have pushed.

Hosting the Streamlit Application in Azure Container Instances

In this section, we will work with the Image pushed to the Azure Container Registry. We will create a Container Instance service and host the Streamlit Image in the ACR. To create an Azure Container Instance, either search in the search bar or directly click on the Container Instance Icon in the dashboard and then click Create Container Instance Button.

Now, we must select the Resource Group, give our Container a name, then set the Image Source to Azure Container Registry. Select the Resource Group that we have created in this article. Here then we can select the streamlit Image that we have pushed to the ACR. Let remaining configurations be set to default. Then click on next.

If you get an Admin-related error, run the command below in your command line. Replace <acrName> with the name that you have used to create the Azure Container Registry.

az acr update -n <acrName> --admin-enabled true

After clicking on Next, we will enter the Network Configuration. Here we see that PORT 80 is open, hence our application will be able to Run. Now if we want, we can provide a Domain Name. After this, keep clicking on Next, until we reach the Create Button. Finally, click on Create to create the Container.

After clicking on Create, we get to see Deployment in Progress status. And after a few minutes, the deployment will be complete. When it’s completed, we can click on the Go to Resource to access the link to our running Container.

Now we see the IP address and the FQDN. Copy one of these and paste it into a new tab in the browser.

Now finally, we can view the application. We see that the application is working properly. Finally, we have successfully hosted a running Container in the Azure Container Instance. The same steps can be used to deploy the Container of your choice through the ACI service.

Conclusion

Azure Container Instance is one of the services provided by Microsoft Azure to host running Containers in the cloud. This is considered Caas (Container As A Service). ACI and Azure Container Registry ACR provide a clean and straight approach to hosting our images in the cloud. And even the time spent on the command line is far less, and most of the work was done through Azure Portal, making things much easy. In this article, we have gone through the process of hosting a website application made through streamlit by creating a running Container in the Azure Container Instances.

Some of the key takeaways from the article include:

• The pricing of Azure Container Instance is comprehensive.
• Deploying through Container Instance is easier because it involves less time spent with the command line.
• The Containers hosted on ACI can either be accessed publicly or privately.
• Provides the users to give custom FQDN.

The media shown in this article is not owned by Analytics Vidhya and is used at the Author’s discretion.

Ajay

I work as a Developer in the field of Data Science. I constantly spend time learning new things be it related to AI, DataSceine, and CyberSecurity. Deep learning and machine learning are two topics that I find particularly fascinating, and Python is my preferred language for programming. Cyber Security is another field that I'm touching upon recently. I have experience with large-scale data analysis, and I have a solid grasp of a variety of deep learning and machine learning approaches, including neural networks, regression models, and natural language processing. I'm eager to take on new challenges and make a meaningful contribution to the industry, so I'm constantly seeking for ways to enlarge and deepen my knowledge and skills in the subject.