Understanding Your Azure Bill
June 24, 2015
Recently ranked by Gartner as an industry leader due to its ability to help organizations do more, work faster and save money, Microsoft Azure is a great, reasonably priced cloud offering. However, the a la carte service pricing model can add up quickly if you don’t properly manage your Azure implementation. In addition to planning your cloud environment to take advantage of Azure’s sizing and scaling capabilities, it’s also important that you review and monitor your monthly Azure usage and invoice in order to understand where you’re spending your money.
The Azure Billing Portal
It’s important to review your Azure bill on regular basis so that you have an idea of what usage charges to expect. You are probably already reviewing the PDF version of your Azure invoice and have found that the invoice is an aggregated summary. The lack of details regarding the instance breakdown and the normalization of Cloud Services and Virtual Machine Sizing can make this process confusing. Fortunately, the Azure Billing Portal also provides an option to download the details of your statement as a CSV file. The CSV extract shows the actual usage by instance. Unfortunately, the daily usage section does not contain billing rate information, which can make detail analysis time consuming.
Creating a PivotTable Report
Your review should be a two-step process. The first step is to view the summary detail in Excel and create a PivotTable report on the name versus the sum of the value (see Microsoft’s Pivot Tutorial). Typically, you’ll find that Azure App Services, Virtual Machines, and Data Services consume most of your budget. Regardless of the actual split, the initial analysis will provide you with a good idea of what parts of your bill you should be concerned about.
Looking At What Matters Most
The next step is to drill into the details to see if they match your expectations. This can be done by creating a pivot on the “Details” section of the extract. In order to create the pivot and have the data aggregate correctly, you will need to remove the date from the component description. A good starting pivot would be to view “Name and Component” by sum of the consumed units.
One “gotcha” when reviewing the usage of Cloud Services and Virtual Machines instances is that the consumed units are calculated as Single Processor Instances. When reviewing the details, some items to inspect are:
- Review the services and instances that are incurring charges
- Validate that all the services that you are charged for are actually in use
- Keep in mind that many items that are billed by the hour are billed regardless of activity, even if they are turned off
- Review the units that you are being billed for
- For items billed by the hour, compare the billed units to the number of hours in the month to determine the average processor instances used in the month. This number can be helpful when reviewing the sizing and scaling of services in your Azure environment.