Investigate without writing queries

Dashboards show what someone already thought to watch. When something new breaks, you need to find out what just changed, and no saved page covers it yet.

Drilldown is guided troubleshooting. You start from Metrics, Logs, Traces, or Profiles, then click filters to narrow down, without writing a query.

Metrics, logs, traces, and profiles

Drilldown has a section for each:

• Metrics: Numbers recorded over time, for example CPU or request rate
• Logs: Timestamped text records of what happened
• Traces: A map of one request as it hops between services
• Profiles: Function-level resource consumption, such as CPU and memory

The same problem often shows up in more than one of these. A slowdown might appear as a spike in the metrics, an error line in the logs, and one slow hop in the trace.

See for yourself

The guide below starts you in Metrics, has you click a filter, and shows the result narrowing. It runs on Grafana Play, where metrics are already available, so you don’t need an account.

To go deeper on investigating your own metrics and logs this way, refer to the Investigate metrics and logs with Drilldown journey.