Reducing Azure VM Costs with Spot Instances and Auto-Shutdown

Overview and What You Will Learn

In this lab, you will configure auto-shutdown on a development VM so it stops billing outside business hours, then create a Spot VM and observe its discounted pricing and eviction behavior - two genuinely different cost-reduction techniques for two genuinely different situations.

Why This Matters in Production

A forgotten test VM at a startup runs 24/7 for months because nobody remembered to turn it off after business hours, quietly costing as much as a small production workload for a resource that was only ever used a few hours a day. Auto-shutdown alone can cut that VM's compute cost by more than half, simply by not billing for the two-thirds of the day nobody was using it.

Core Principles

Auto-shutdown and Spot Instances solve two different cost problems and should not be confused with each other.

Auto-shutdown is appropriate for any VM with a predictable, non-24/7 usage pattern. Spot Instances are appropriate specifically for workloads that can tolerate being interrupted and restarted elsewhere - never for anything customer-facing that needs to be reliably available.

Detailed Step-by-Step Practical Lab

1. Create a Resource Group and a development VM:

2. Configure auto-shutdown on the dev VM to power off automatically every night at 9 PM:

3. Confirm the auto-shutdown schedule was applied:

4. Create a separate Spot Instance for a batch-processing workload, accepting Azure's spare capacity at a steep discount:

Note

--max-price -1 means you're willing to pay up to the current on-demand price rather than setting a hard cap - this is the simplest configuration to start with. --eviction-policy Deallocate means an evicted Spot VM is stopped and deallocated rather than deleted, so its disk and configuration are preserved for when it can be restarted.

5. Confirm the Spot VM was created with the expected pricing model:

6. Review Azure Advisor's cost recommendations to see if any additional rightsizing opportunities exist on VMs already running:

7. Clean up:

Production Best Practices & Common Pitfalls

Common Mistake

Using a Spot Instance for a customer-facing or otherwise critical workload to save money, then experiencing an unexpected outage when Azure reclaims the capacity with only a couple of minutes' notice. Spot pricing is a genuine discount, but it comes with a real availability trade-off that only fits fault-tolerant, interruption-tolerant workloads.

Tip

Combine auto-shutdown with a corresponding auto-start schedule (via Azure Automation or a Logic App) for dev/test environments used on a predictable daily pattern, so the environment is ready and running again automatically each morning without anyone needing to start it manually.

• Auto-shutdown only helps VMs with genuinely predictable idle time. A production VM serving live traffic around the clock gets no benefit from auto-shutdown, since it's never actually idle in the first place.
• Reserved Instances are the right tool for stable, predictable, always-on production workloads - a separate cost-reduction lever from both auto-shutdown and Spot, appropriate when you know a VM size will run continuously for the next 1-3 years.

Quick Reference & Troubleshooting Commands