Azure Cosmos DB is great at scaling, but when something starts behaving strangely—latency jumps, RU consumption suddenly spikes, or your application gets flooded with 429s—figuring out the real cause isn’t always obvious. The symptoms are visible, but the underlying issue often isn’t.

In this session, we’ll explore how to use Azure Cosmos DB diagnostics to investigate and understand these kinds of issues. Instead of guessing, we’ll look at the data Cosmos DB already provides and learn how to read it properly. You’ll see how diagnostics can reveal things like inefficient queries, hot partitions, unexpected retries, and other hidden performance problems.

Using a few real-world troubleshooting scenarios, we’ll walk through how to move from “something is wrong” to identifying the real cause behind the behavior. The goal of this session is simple: help you feel more confident when troubleshooting Cosmos DB workloads and give you a set of techniques you can immediately apply in your own systems.