Azure Spot VMs for cost savings - Step-by-Step Execution

Process Flow - Azure Spot VMs for cost savings

Request Spot VM

↓

Check Available Capacity

Yes↓

Allocate Spot VM at Discounted Price

↓

Run Workload

↓

If Capacity Lost or Price Exceeds Max

Yes↓

Evict Spot VM

↓

Option: Redeploy or Stop

Back to Request Spot VM↓

Wait or Use Regular VM

↩Back to Request Spot VM

The flow shows requesting a Spot VM, checking capacity and price, running workload, and handling eviction if capacity or price conditions change.

Execution Sample

Azure

Request Spot VM with max price $0.05
If capacity available and price <= max
  Allocate Spot VM
Run workload
If eviction signal received
  Evict VM

This code requests a Spot VM at a max price, runs workload if allocated, and evicts VM if Azure reclaims capacity.

Process Table

Step	Action	Capacity Available?	Current Spot Price	Max Price	VM Allocated?	Eviction Signal	Result
1	Request Spot VM	Yes	$0.04	$0.05	No	No	Check capacity and price
2	Evaluate price and capacity	Yes	$0.04	$0.05	Yes	No	Allocate Spot VM
3	Run workload	Yes	$0.04	$0.05	Yes	No	Workload running
4	Price rises to $0.06	Yes	$0.06	$0.05	Yes	No	Price exceeds max, eviction imminent
5	Eviction signal received	Yes	$0.06	$0.05	Yes	Yes	Evict Spot VM
6	Redeploy or stop	No	$0.06	$0.05	No	No	VM stopped or redeployed
7	Request Spot VM again	No	$0.06	$0.05	No	No	Capacity unavailable, wait or use regular VM

💡 Execution stops when capacity is unavailable or VM is evicted due to price or capacity changes.

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	After Step 5	After Step 6	After Step 7
Capacity Available	Unknown	Yes	Yes	Yes	Yes	Yes	No	No
Current Spot Price	Unknown	$0.04	$0.04	$0.04	$0.06	$0.06	$0.06	$0.06
Max Price	$0.05	$0.05	$0.05	$0.05	$0.05	$0.05	$0.05	$0.05
VM Allocated	No	No	Yes	Yes	Yes	Yes	No	No
Eviction Signal	No	No	No	No	No	Yes	No	No

Key Moments - 3 Insights

Why is the Spot VM evicted even though capacity is still available?

Because the current Spot price rose above the max price set by the user (see Step 4 and Step 5 in execution_table), Azure evicts the VM to reclaim capacity.

What happens if capacity is not available when requesting a Spot VM?

The VM is not allocated (Step 7 in execution_table). The user can wait or choose a regular VM instead.

Does the VM run continuously once allocated?

No, the VM runs until Azure sends an eviction signal due to price or capacity changes (Step 3 to Step 5).

Visual Quiz - 3 Questions

Look at the execution_table, at which step is the Spot VM first allocated?

A Step 1

B Step 2

C Step 3

D Step 4

💡 Hint

Check the 'VM Allocated?' column in execution_table rows.

According to variable_tracker, what is the 'Current Spot Price' after Step 4?

A $0.04

B $0.05

C $0.06

D Unknown

💡 Hint

Look at the 'Current Spot Price' row under 'After Step 4' in variable_tracker.

If the max price was increased to $0.07, what would happen at Step 4 in execution_table?

A VM would continue running as price is below max

B VM would be evicted due to price

C Capacity would become unavailable

D Eviction signal would be ignored

💡 Hint

Compare 'Current Spot Price' and 'Max Price' columns in execution_table at Step 4.

Concept Snapshot

Azure Spot VMs let you use spare capacity at a discount.
You set a max price you are willing to pay.
VMs run until Azure reclaims capacity or price rises above max.
Evicted VMs stop or can be redeployed.
Best for interruptible, flexible workloads.

Full Transcript

Azure Spot VMs provide cost savings by using spare capacity at discounted prices. You request a Spot VM with a maximum price you agree to pay. If capacity is available and the current Spot price is at or below your max price, Azure allocates the Spot VM. Your workload runs on this VM. However, if Azure needs the capacity back or if the Spot price rises above your max price, Azure sends an eviction signal and the VM is stopped or redeployed. If capacity is unavailable when requesting, the VM is not allocated and you can wait or use a regular VM. This approach is ideal for workloads that can handle interruptions and want to save costs.