Getting started with Azure Cloud Solutions

Azure Cloud Solutions are the most in demand solutions in the Tech Consulting space right now. Most organizations that have already been using Microsoft 365, Sharepoint files, SSRS and Microsoft DBs, it's easier to sell cloud services to them since they are already Microsoft Users - and who does not want to be left out?

With Companies needing to go virtual as experienced from the Pandemic, there is now realisation that cloud is indeed a good technology to use.

But how do we get started with Azure Cloud? a lot have uncertainty to Cloud solutions aside from Devs entering their credit card numbers and get a shocking monthly bill

Well I tried it on my own using Azure Cloud services and they do have a lot of services available that you can enable. I started using Azure Machine Learning both on Jupyter notebooks and their AutoML. Needless to say they also offer the $200 credits that you can use to play around and get used to their UI. It's got more features now when they already include an Inference Cluster (this is what you use to deploy your ML algorithm)

Compute Power

I wanted to use the real good compute such as the compute cluster (these are the ones with more than one node) but if you are cheap, you can always use the compute instance with single nodes. (think like a laptop machine somewhere where you can put it on for 24 hours) but the problem with this is that it's not super powerful and when you run your algorithms it might take more than a couple of minutes. (Cheap but takes long to finish)

Storage

Now, let's go to storage, this is the 2nd most high variable on my billing. Well it's actually credits but I did use some storage space, creating duplicate tables here and there and storing it on AWS BLOB storage (it can both house structured and unstructured data, and you can even use Data stores for it) there are specific paths that you can save on your notepad if you want to track them down. Or just go to your storage folder and check the last of your data sets. I did not delete the resource, since it only took me around $5 for the CSV files that I loaded for bank churn data.

Network

I created two subscriptions, one in North America East 2 region and another in the Singapore region. I created these two regions just to play around. But it did cost me more since I was working with two regions instead of one. The rule of thumb is to stick to just one region and put everything there - and that seems practical and logical. Put your storage, your compute and your apps in one region instead.

Billing Alerts

So far these are the 3 main things that I try to track on my monthly billing specific to Microsoft Azure Cloud. If you want to set up a billing alert, you can put up a specific threshold and this will go directly to your e-mail inbox.

Deleting Resources

Now, if you are still cheap and still can't get a good night's sleep, you can always delete your resources. Delete first the Virtual Machines or your Compute Cluster once you don't need them. Next, I deleted the data sets or the storage that I used up. Then finally I delete the subscription tied to it and come up with a clean slate.

Free Services

Microsoft Azure does give a list on Free Services, on your 12 months, some of them are

• Azure Cosmos DB for 25GB storage
• Azure App storage for 1 GB storage
• Azure Functions for 1M requests
• Azure Event Grid for 100k operations
• Azure Kubernetes Service
• Azure Data Factory for 5 low frequency activities
• Azure API management for 1M monthly calls

----------------------------------------------------------------------------------------------

About the Author

Bert has 22 years work experience and has an MBA degree from De La Salle University and a bachelor's degree in Mathematics specializing in Statistics and Actuarial Science. He is a book author on Google Play: A Random Walk by Albert Gavino. He is also the administrator of Data Science Manila FB page with around 8k members as of date.