Should I sign a three-year cloud service contract?

Buddy, we need to think this through carefully, don't be impulsive. Signing a three-year contract is a bit like renting an office for three years – it has its pros and cons.

First, let's talk about the benefits, which is why cloud providers want you to sign long-term agreements:

It's cheaper! This is the biggest temptation. Signing for three years usually comes with the largest discounts, potentially 30% or even more compared to monthly payments. For a startup, every penny saved counts.
Peace of mind. The price is locked in, so you won't have to worry about price increases for the next three years, and your budget will be very clear.

However, the risks and pitfalls, especially for a startup, are much greater:

Are you really sure what you'll need in the next three years? The biggest characteristic of a startup is "change." You might think you need a server of a certain size today, but three months later, if the product direction shifts, those resources might be completely insufficient; or the project might fail, and the server sits idle, burning money for nothing. You're locked into the contract; adding capacity might be very expensive, and reducing it? Sorry, no refunds.
Technology updates too fast. You might think this cloud provider is great now, but a year from now, a better, cheaper, or more suitable one might emerge. At that point, you can only watch enviously because your contract hasn't expired. This is called "vendor lock-in."
"Vendor lock-in" is deadly. Once you use a provider's service, your code, data, and operational practices become deeply tied to them. Want to switch to another provider three years later? That could be a massive undertaking, with migration costs so high you won't want to move.
What if the company fails? This isn't pleasant to say, but startups are a matter of survival of the fittest. If the company doesn't make it past three years, you might still be stuck with the cloud service contract debt.

Therefore, my advice is:

For a startup, absolutely, absolutely do not easily sign for three years! Flexibility is far more important than saving money. Cash flow and the ability to pivot are your lifelines.
Start with "pay-as-you-go." Pay for what you use; this is the most flexible option. Although the unit price might be higher, you're not wasting anything, and the risk is minimal.
Sign for a maximum of one year. Once your business is up and running, and you have a rough idea of your resource needs, you can consider a one-year contract to get a discount. One year is enough time to validate your business model and gives you the opportunity to re-evaluate next year.
When can you consider a long-term contract? Only if your business is extremely stable, for example, if you're just a corporate website and your traffic won't change much in the next three years. In that case, you can sign a long-term contract to lock in that cost. But for anything related to core products or user data, be very cautious.

In summary: For startups, flexibility is king. Don't tie your hands for three years just to save a little on discounts. Get started, survive, that's more important than anything.