Azure database savings plans: the 3-year term isn't about a bigger discount

Savings plan for databases picked up a second term option this year: alongside the original 1-year commitment, you can now choose 3 years. If you work with Azure compute savings plans or reservations, the instinct is to assume a longer term buys a deeper discount, the way it usually does elsewhere. For savings plan for databases, that assumption is wrong, and Microsoft says so directly.

Same discount, different question

Microsoft’s announcement is explicit: both terms provide the same discounted pricing. The choice isn’t about finding a bigger number, it’s about matching the commitment period to how stable your database footprint actually is.

That reframes the decision entirely. A 3-year term makes sense when you already know a workload is staying put for a while and you’d rather not revisit the commitment every 12 months. A 1-year term makes sense when you expect to keep reassessing, whether that’s because the workload is still evolving or because you just prefer the shorter leash. Neither one is the “better rate” option, because there isn’t one.

Why this is easy to miss

It’s an easy assumption to carry over by accident. Azure compute savings plans and reservations both reward longer commitments with meaningfully better rates, so “3-year beats 1-year on price” is a reasonable prior if databases are the newer plan you haven’t looked closely at yet. The database plan breaks that pattern on purpose: it’s designed for teams whose architecture keeps changing (services, regions, configurations) even when their overall spend is predictable, so the flexibility is in what the commitment covers, not in trading term length for rate.

What to actually check before you buy

Since the discount rate isn’t the deciding factor, the useful question is planning-horizon, not pricing: how confident are you that this database spend looks the same in year two and three? Azure Advisor and the savings plan purchase experience in the Azure portal both surface commitment recommendations based on your actual eligible usage, which is a better starting point than reasoning about it in the abstract. If a 3-year commitment doesn’t feel right for a given workload, the 1-year term costs you nothing but an earlier renewal conversation.

Deciding on a term is a small piece of a bigger commitment strategy. If you want a second set of eyes on where reservations, savings plans, and database plans should each fit into yours, that’s exactly the kind of thing a workshop session is built for.