Azure

1. What is Azure? 

• Azure is a online cloud computing plateform created by Microsoft to build, test, deploy and manage applications and services.
• The applications and services gets deployed on Microsoft managed datacenters.  

Here are following purchasing Options:

• Pay as you go (PAYG) - Pay for the services you use. 
• Enterprize Agreement (EA) - It re-quires a commited amount you woud spend in Azure, Paid upfront. 
• Dev/Test - It is discounted Azure service only when Development and testing need to be done. 
• Cloud service provider (CSP) - In this case, Microsoft partner sell services to customer. 

2. What is Azure reservation? 

• Purchasing a long term service upfront (1 to 3 years) 
• Discount varies based on term of commitment 

3. Product Pricing (Estimate)

• Virtual Machines The price chagnes based on the region, operating system due to licensing cost. Basic tier can be used for development and testing as it doesn't use load balancer, scaling feature etc. 
• Kubernetes - You pay for node which perform tasks not for cluster.  
• Azure Function - You pay only for execution time. 
• App Service - Various options like Basic, Standard, Premium etc. 
• For Dev/Test - F1 and B1 series machies 
• For Producton - P1V1, P1V2 & P1V3 series available. 
• Storage Account - Price varies based on capacity (How much we store) and access tier (Hot, Cold or Archive). In case Hot tier, One pay more for Capacity then Read or Retrive.

Scenario: 

• Hot tier is cheapest in case you have too many read and write operation.
• Archive tier is cheapest in case you have to store/archive the data with limited access.

4. How do we know a resource like VM box is underutilized?How cost can be optimized?

Azure advisor is a build-in service that can be utilized to identify a resource that is underutilized. Azure advisor recommendations migh include:

• Resizing or shutting down underutilized instances.
• Deleting unprovisioned ExpressRoute circuits.
• Deleting or reconfiguring idle virtual network gateways.
• Azure Reservations for virtual machines.
• Deleting unassociated public IP addresses.
• Deleting Azure Data Factory pipelines that are failing.
• Using standard snapshots for managed disks.