What the waterfall told you

You just opened a real trace and read it yourself, start to finish, with no setup. Reading a trace and understanding what it tells you is the core skill, and it transfers to any trace on any surface.

Here’s what you pulled out of the span waterfall:

• Took in the trace’s shape at a glance.
• Followed the parent and child nesting to see which call ran inside which.
• Read the durations and found where the request spent its time.
• Learned to open a span and read its attributes to tell what that step did.

Two doors open from here:

• A query language for finding traces on purpose. You ran a ready-made TraceQL query to land on your practice traces. When your questions get specific, like “failed checkout requests slower than two seconds,” TraceQL is how you pin them down.
• Connecting a trace to your other signals. A trace is even more useful next to the logs and metrics from the same request. You’ll get a lot more out of a trace once you connect it to those signals, but you have to set that up first.

Related paths

Consider these paths after you complete this journey:

• Use Traces Drilldown to identify high latency service requests: investigate your own traces with a queryless workflow
• Send traces to Grafana Cloud using Alloy: get your own traces flowing
• Visualize traces in a Grafana Cloud dashboard: build trace dashboards and learn TraceQL

More to explore (optional)

• Trace structure
• TraceQL cookbook: copy-paste query recipes to find traces on purpose
• Traces in Explore
• Traces Drilldown