Moving Beyond Monitoring: Why Your App Needs True Observability

Hey fellow devs!

Happy New Year! We’re kicking off the first session of the year with a deep dive into something that’s becoming absolutely essential for every backend engineer: Observability. We often use the terms “monitoring” and “observability” interchangeably, but as we discussed in our latest session, they are very different animals.

Monitoring vs. Observability: What’s the Catch?

In the old days, we were mostly worried about basic metrics. “Is the server up?” “How’s the CPU looking?” This is what we call Monitoring. But as our codebase grows and we move into the world of microservices, just knowing the server is “up” isn’t enough.

Why “Up” is No Longer Enough

Imagine you have hundreds of interconnected services. A request comes in, and suddenly there’s an error. Where did it happen? Which service timed out? This is where observability comes in. It’s about understanding the behavior of your application from the inside out.

As Thiago Gonzaga put it during our talk:

Thiago: It’s not just about monitoring. In the past, we used the term monitoring because we just saw if the application was up or not, or checked basic metrics like CPU and memory. But nowadays, we need to understand much more. We need to understand the path our user takes and how this tree opens up in our system.

The Three Pillars: Metrics, Logs, and Traces

To get true observability, you need to correlate three types of data:
1. Metrics: The numbers (latency, error rates, etc.).
2. Logs: The detailed story of what happened.
3. Traces: The path a request took through all your services.

When you match these three, you can find the root cause of a bug in minutes instead of hours. You can see exactly which database query slowed down a specific user request.

Wrapping Up

The goal is to ensure a smooth experience for our users. We can’t do that if we’re flying blind with just a “server up” light.

Have you started using tracing in your microservices yet? Let me know in the comments below!

Happy coding!

Cheers,

Pedro Cavalero