Reserved Instances vs Spot Pricing Explained

Azure offers several purchasing models for compute, and the exam consistently tests whether you can match a workload's characteristics to the pricing model that actually fits it — not just recall the discount percentages.

Pay-as-you-go is the default, no-commitment model — you pay only for what you use, with the flexibility to start or stop anytime. It suits unpredictable or short-term workloads where you can't commit to steady usage in advance.

Reserved Instances offer a significant discount — commonly cited as up to around 72% compared to pay-as-you-go — in exchange for committing to one or three years of consistent usage. This model suits steady, predictable workloads: a production database that runs 24/7 for years is the textbook case.

Spot pricing offers the deepest discount available, sometimes up to around 90% off pay-as-you-go rates, but with a critical trade-off: Azure can evict your Spot virtual machine at any time when it needs the capacity back for higher-priority workloads. This model only suits interruptible workloads — ones that can tolerate being stopped and restarted without causing a problem, like batch rendering jobs or fault-tolerant background processing.

The exam trap to watch for: a scenario describing a workload that absolutely cannot be interrupted (a customer-facing production application, for example) but the tempting wrong answer is Spot pricing because of the large discount. Cost alone is never the deciding factor — the workload's tolerance for interruption is.

Azure Hybrid Benefit is a separate, complementary saving — it lets organisations apply existing on-premises Windows Server or SQL Server licences (with Software Assurance) to Azure VMs, and can be combined with Reserved Instances for additional savings on top.