Azure Savings Plan for Compute
What is it?
Azure Savings Plan for Compute is a new pricing model where eligible customers commit to spend an hourly amount of money on select Azure compute services in exchange for lower prices.
The plan provides a convenient and flexible way to save up to 65% on the cost of compute services compared to pay-as-you-go prices. The plan can apply over a one-year or three-year period, depending on your choice. You can also pay the total commitment amount upfront or monthly at no extra cost.
Unlike Azure Reservations, you can scope the savings plan commitment across multiple Azure regions.
The following compute services participate in the savings plan: (please note only a few eligible Azure services are listed.)
- Azure Virtual Machines (same VM SKUs eligible for Azure Reserved Instances)
- Azure App Service (Premium v3 and Isolated v2)
- Azure Dedicated Host
- Azure Container Instances
- Azure Functions Premium plan
Why Buy a Savings Plan?
If you have consistent compute spending across multiple regions, buying a savings plan allows you to reduce your costs. For example, when you continuously run instances of a service without the savings plan, you're charged the pay-as-you-go rates. When you buy a savings plan, your compute usage is immediately eligible for the savings plan discount. Your discounted rates add up to the total commitment amount on the contract.
When Should a Customer Use Reservations vs. Savings Plan?
Both options provide long-term savings for customers' compute usage, and it is best to align the offer to the workload pattern. Reserved instance provides the most significant cost savings for stable, predictable workloads with no planned changes, and Azure Savings Plan for Compute is better suited for workloads that are dynamic and in need of greater flexibility.
Recommended use cases for Azure Savings Plan for Compute
- Global Compute or follow-the-sun Compute models leveraging Virtual Machines, Dedicated Hosts, Container Instances, App Services, and Premium Functions in any region
- An Example of this would be organizations that leverage Azure Virtual Desktop Globally.
- Dev/test workloads that get powered on and off consistently or are not predictable otherwise.
- Flexibility in purchasing and configuring your Azure compute architecture regardless of region, instance series, or operating system.
- An efficient way to support dynamic and evolving workloads or transform compute environment with planned or unplanned changes.
The Advantages of Azure Savings Plan for Compute
It allows customers to commit to a set hourly spend on selected Azure compute services, while Azure Reserved Instances require a specific commitment to a particular virtual machine.
You can save up to 65% savings compared to pay-as-you-go prices, which can result in significant cost savings for eligible customers.
Pay the total commitment amount upfront or monthly at no extra cost, which provides customers with more payment flexibility.
Your commitments can be scoped across multiple Azure regions, making it easier for customers to manage their costs across a multi-region deployment. You can only apply Azure Reserved Instances to a specific region.
The Advantages of Azure Reserved Instances
Azure Reserved Instances commit to a specific virtual machine, allowing customers to control and predict costs.
Azure Reserved Instances provide a predictable and stable cost model, which can benefit customers with predictable workloads.
Azure Reserved Instances offer the option to pay upfront for the total commitment, which can result in significant cost savings for customers with longer-term needs.
Azure Reserved Instances can help customers with capacity planning as they provide a guaranteed capacity for a specific virtual machine, making it easier to manage resource utilization.
What should customers consider as they prepare to set an hourly commitment amount?
Customers can review their personalized recommendations in the Azure portal based on their recent usage. It's essential that they also consider future needs 1 or 3 years out when setting the hourly commitment amount.
You should evaluate the optimal hourly commitment after savings are applied using lower saving plan prices vs. pay-as-you-go.
Can Customers Cancel or Modify a Savings Plan?
Savings plans cannot be modified or canceled once you have committed spending. If their usage needs have grown beyond the current savings plan, customers may add another savings plan to cover the additional usage.
How is a Savings Plan Billed?
At the time of purchase, customers can pay for the savings plan in a single upfront payment or divided into equal monthly US dollar payments at no extra cost.
Key differences between Azure Savings Plan for Compute versus Azure Reservations:
Azure Savings Plan
- Options available to scope across regions, subscriptions, and resource groups
- Cannot be exchanged or returned
- A single savings plan can cover multiple types of compute, such as multiple virtual machine families, App Services Premium, and Azure Functions Premium.
- Options available to scope across subscriptions and resource groups but not regions
- It can be exchanged until January 1st, 2024. Please refer to the Azure Reserved Instance Exchange Policy document for more information: Azure Reservation Exchanges.
- A single Reservation only applies to a single type of compute, such as a single Virtual Machine Family.
- When Microsoft calculates billing, they utilize Azure Reservations first, then apply Azure Savings Plan for Compute commitments to remaining eligible resources. The rest of the Azure utilization is calculated at pay-as-you-go rates.
- We recommend determining which workloads best fit Azure Reservations first, then choosing the appropriate commitment amount for Azure Savings Plan for Compute to maximize your savings.
- Each eligible Azure compute resource type (down to the Virtual Machine family) may have a different discount percentage within the same region and even in different regions, including a different discount rate for 1-year versus 3-year terms.
- Azure Savings Plan for Compute commitments can be "stacked," such that you could start with a single commitment, then add another commitment that overlaps (partially or fully) the scope of the first one before the term of the first commitment has expired.
- Azure Advisor provides estimated savings for both Azure Reservations and Azure Savings Plan for Compute.