Tempo

The third leg of observability are traces.

They help you track requests across service boundaries, allowing you to trace your customers experience as their requests travel through the architectures of your microservices.

It is the least used of the three approaches, and it also requires an upfront investment to propagate request headers through your services.

The promised land

But once you do the header propagation and ship traces to Tempo, the overview they provide nicely augments your troubleshooting workflows.

[TODO] Gimlet Stack packaged traffic source and screenshot of the service map

How requests are identified

[TODO] Show how to extract the traceID from traces.

How to get started

Since traces need upfront work, it is easier to go through this process, if you have a traffic source that simulates how things look when everything set up.

Gimlet Stack has an installation method for Tempo with a packaged traffic source, so you can judge the trace experience for yourself.