Today, instrumenting applications to observe their performance and behaviour using OpenTelemetry is easier than ever. However, if you are an engineer tasked with instrumenting applications with OTel, it’s still not a quick job. Things that still make this non-trivial:

– Your application consists of multiple services? You want to have the complete picture? Hm, then you really ought to instrument it all. But OK, it can be done incrementally.

– Some of your applications are in Python, some in Go, maybe some Java? Each of them requires slightly different instrumentation.

– Oh, you want to see truly end-to-end transactions? Well, then you need to instrument your frontend, too, not just the backend. Oh, and make sure the context/trace ID is passed from the frontend to the backend.

We experienced this ourselves when we started instrumenting our own services. To make it easier for ourselves and for other Sematext users to instrument applications with OpenTelemetry SDKs we’ve published an AI Skill you can use with Claude Code or any of your AI friends.

The skill lives in Github. We published a short guide for this skill, along with the accompanying blog.

When you instrument your applications and services you will immediately see all your services and infrastructure as a Service (Dependency) Map, deep insight into your application performance, errors, etc., ability to drill into the root cause of synthetic monitor failures, and more.

Start Free Trial