Is Serverless a good idea for an insurtech startup ?

At Dalma , we decided from day one to go fully serverless.

Four years later, we’re managing 341 Lambda functions, 77 state machines, and 50 DynamoDB tables.

Whenever we present our architecture, we often hear the same reactions:

• “You’re crazy to go serverless. Serverless is just for prototypes.”
• “There’s no way you can maintain that many functions.”
• “Your performance is going to be terrible.”
• “Your costs will skyrocket. You need to get out while you can!”
• “You won’t be able to sustain this architecture as your team grows.”

“You’re crazy to go serverless.”

Not a big surprise here. We’re always looking for innovative technologies that solve real problems. But here’s the thing: we’re not just “serverless”—we’re full serverless.

Serverless isn’t about simply combining Lambda, Step Functions, and other trendy services. It’s about finding the right architecture that minimizes both maintenance overhead and costs, while solving the problem at hand.

“It’s impossible to maintain that many functions.”

Maintaining this many functions is challenging, but definitely not impossible. The key lies in designing the right architecture.

To make it work, you need a deep understanding of the problem you’re trying to solve. This means investing time and energy into understanding both the current needs and future requirements.

At Dalma, we still have Lambda functions that were last modified over two years ago—and they’re still working perfectly.

Even if they don’t follow our most recent development conventions, we’re not planning to replace them as long as they continue to solve the problem.

Even if they are not following our latest development conventions, we are not planning to replace them as long as they solving the problems.

“Your performance will be terrible.”

The sky is blue, oranges are orange, and our performance is just fine. 😊

One of the main advantages of our serverless setup is auto-scaling, which helps us handle spikes in demand and multiple simultaneous invocations.

Plus, we’ve designed our architecture so that each Lambda function has a very specific task to perform and executes it as efficiently as possible.

“Your bill will be enormous.”

If you understand how AWS billing works, it’s actually quite straightforward to keep costs under control by choosing the right services and optimizing your orchestration.

The worst mistake is to build first and then check the bill. That’s how you end up with a headache.

how much our bill is in 2024?

*Less than $300 per month. Surprised? 😵💫
*
And that’s without fully optimizing our architecture yet. We believe we can reduce this bill by another 3x.

“You won’t be able to maintain this architecture as your team grows.”

Yes, scaling this architecture as the team grows is definitely a challenge. Not everyone is comfortable with or understands the nuances of serverless.

But fortunately, we’ve been lucky to find extraordinary people who have helped us push this architecture far beyond what we expected.

Conclusion:

Going full serverless was a bold choice, but it’s paid off. It hasn’t been easy, but it’s been incredibly rewarding.
The key to making it work lies in solid architecture, a deep understanding of the problem, and a commitment to continuously optimizing.

At Dalma, we’re proof that full serverless can scale, perform, and be cost-effective—if done right.