This week I got some inspiration to productionize a piece of work I did for a client a few years back. This was the idea of making sure all your tests are traces.
Why is this useful? The inner dev loop is important to fast and safe software delivery, and what better way to make that safe than for you to understand how a system works from its tests?
In this NuGet package, I’ve made it so that each test is its own trace–but further, any spans that your application code make become part of that same trace, so now you can see things like why a test is slow.
This can show you all sorts of things before you think about deploying, but also when you’re debugging.
Check it out at https://github.com/practical-otel/dotnet-xunit-otel
