Is Microsoft Azure Cosmos DB Really Fast?

When building high-load systems, one of the biggest challenges is making sure your solution performs well under pressure. This is especially important when your system relies heavily on a database — often the main bottleneck for throughput. That’s exactly what happened in one of my recent projects, where Cosmos DB turned out to be the limiting factor. The client had strict performance requirements, and meeting them meant I had to dive deep into what Cosmos DB is really capable of — and how to get the most out of it.

At first, I didn’t have a clear answer to a simple but critical question: how fast can we actually send data to Cosmos DB? Microsoft’s documentation gave some general ideas and high-level guidelines, but it wasn’t enough to design a solid, performance-focused solution. To make things more challenging, the client chose Kotlin as the development language — and there’s no official Cosmos DB SDK for Kotlin.