Understanding the savings on a schedule

This article walks through a configured schedule example and explains how savings are calculated and realized.

Top Level Schedule

The table below defines the key columns at the schedule level.

Column	Description	More Info
List Cost	This is the sum of the list price for resources within the schedule	This is based on a 30 days cost
Potential Savings	This is the sum of the potential savings at resource level
Amortized Cost	This is the sum of the amortized cost for the resources in the schedule	This is based on the last 30 days This will include any impact of savings plans and reservations or hybrid benefit

Schedule Time Settings

The schedule defines the daily start and stop times for its resources.

This equates to 50 hours of uptime and 118 hours of downtime.

If the schedule is up for 50 hours and down for 118 hours then this means that for >70% of the time the resources will be in the down state and we will hope to see in the region of 70% savings compared to if the resources were in the up state for 100% of the time.

Schedule Resources

This schedule has 3 resources.

Compared to the schedule-level totals in the Top Level Schedule section above:

List Cost: The schedule-level list cost is the sum of the resource-level list costs. An explanation of list cost is below.
Savings: The schedule-level savings is the sum of the resource-level savings shown above.
Amortized Cost: The schedule-level amortized cost is the sum of the resource-level amortized costs. An explanation of amortized cost is below.

The resource configuration shows:

Row in Schedule	Schedule Action	Up Configuration	Down Configuration
1 - App Service Plan	Downsize out of hours	Scale up to P2V3 Premium	Scale down to P0V3 Premium
2 - VM	Turn off out of hours	On (Standard D4as V4 SKU)	Off
3 - VM	Downsize out of hours	Scale up to -> (Standard D32as V4)	Scale down to -> (Standard D2as V4)

The cost-related columns in the resource view show the following:

Column	Source	Description	More Info
List Cost	Azure list price for the sku	This is the Azure list price for the resource if it was in the up configuration for the entire 30 day period.	We use a 30 day period here so there is no confusion between months with different number of days.
Savings	Estimated based on configuration in the schedule	This is the expected savings compared to the list price based on the number of up and down hours configured on the schedule. It is based on a 30 day comparison.	As an example, if the list price for the resource was $100 and you turn it off for 50% of the time we would expect the savings to be $50
Amortized Cost	Based on previous cost data we have collected	This is the amortized cost that you have paid for this resource for the last 30 days	The typical example should be, with a virtual machine on a pay as you go pricing with no savings plans or RIs and if the schedule was to turn the VM off for 50% of the time then the VM which costs $100 should have an amortized cost of $50. There are other factors that can come into play in some circumstances: This will factor in any cost you have paid for any reservations, savings plans which may have been applied to the resource. If you have a reservation partially applying to the resource for different amounts month over month then this may make the amortized cost for the resource change when you look at this value from 1 month to the next. There may be cases where for the month you configure the schedule the Amortized cost is a lot lower than you might anticipate if you are increasing the size of the VM compared to normal

Column

Source

Description

More Info

List Cost

Azure list price for the sku

This is the Azure list price for the resource if it was in the up configuration for the entire 30 day period.

We use a 30 day period here so there is no confusion between months with different number of days.

Savings

Estimated based on configuration in the schedule

This is the expected savings compared to the list price based on the number of up and down hours configured on the schedule.

It is based on a 30 day comparison.

As an example, if the list price for the resource was $100 and you turn it off for 50% of the time we would expect the savings to be $50

Amortized Cost

Based on previous cost data we have collected

This is the amortized cost that you have paid for this resource for the last 30 days

The typical example should be, with a virtual machine on a pay as you go pricing with no savings plans or RIs and if the schedule was to turn the VM off for 50% of the time then the VM which costs $100 should have an amortized cost of $50.

There are other factors that can come into play in some circumstances:

This will factor in any cost you have paid for any reservations, savings plans which may have been applied to the resource.
If you have a reservation partially applying to the resource for different amounts month over month then this may make the amortized cost for the resource change when you look at this value from 1 month to the next.
There may be cases where for the month you configure the schedule the Amortized cost is a lot lower than you might anticipate if you are increasing the size of the VM compared to normal

Points to note

In my example above you may notice that the VM (3rd row) I will be downsizing has a smaller than you might expect amortized cost. This is because in reality this resource has been set as a Standard D2as v4 VM SKU for the last month. In the schedule I will be moving it to a bigger size than normal during working hours so users have more power available but in the evening I will be shrinking it back to the smallest size.

In this sample schedule there are also reservations which apply to these resources quite often which can also affect the amortized cost which is why the amortized cost can look lower than the list cost before you start using the scheduler. In these Azure will attempt to reallocate your reservation to other azure resources provided there are resources which would be within the scope of the reservation. You need to be careful though if you are using a resource on a schedule that you intend to be picked up by a reservation, these are often competing optimization approaches and you need to check you don't end up with underutilized reservations.

Calculation

Resource level savings

Resource	If un scheduled	Scheduled Up	Scheduled Down	Savings
1 = App Service	P2V3 Premium App Service Region = West US 3, Linux 168 hours per week Cost Per Hour = $0.310 Expected Cost per week = $52.08 Expected cost for 30 days = $223.2	P2V3 Premium App Service Region = West US 3, Linux 50 hours per week Cost Per Hour = $0.310 Expected Cost per week = $15.50 Expected Cost for 30 days = $66.42	P0V3 Premium App Service Region = West US 3, Linux 118 hours per week Cost Per Hour = $0.078 Expected Cost per week = $9.15 Expected Cost for 30 days = $39.21	Unscheduled Cost = $223.2 Scheduled Cost = $127.57 Up = $66.42 Down $39.21 Savings = $95.63
2 = VM to stop / start	Standard D4as V4 Region = Central US State = On 168 hours per week Cost Per Hour = $0.217 Expected Cost per week = $36.45 Expected cost for 30 days = $156.24	Standard D4as V4 Region = Central US State = On 50 hours per week Cost Per Hour = $0.217 Expected Cost per week = $10.85 Expected cost for 30 days = $46.50	Standard D4as V4 Region = Central US State = Off 118 hours per week Cost Per Hour = 0 Expected Cost per week = 0 Expected cost for 30 days = 0	Unscheduled Cost = $156.24 Scheduled Cost = $46.50 Up = $46.50 Down $0 Savings = $109.74
3 = VM to scale up/down	Standard D16as V4 Region = East US State = On 168 hours per week Cost Per Hour = $0.768 Expected Cost per week = $129.02 Expected cost for 30 days = $552.96	Standard D16as V4 Region = East US State = On 50 hours per week Cost Per Hour = $0.768 Expected Cost per week = $38.4 Expected cost for 30 days = $164.57	Standard D2as V4 Region = East US State = On 118 hours per week Cost Per Hour = $0.096 Expected Cost per week = $11.328 Expected Cost for 30 days = $48.54	Unscheduled Cost = $552.96 Scheduled Cost = $213.11 Up = $164.57 Down $48.54 Savings = $339.85

Schedule level savings

The total schedule-level savings is calculated by adding up the resource-level savings.

In this case, Total Savings = $545.22