Documentation forSolarWinds Observability SaaS

Exemplars

Exemplars connect aggregated metric data points to specific service traces, helping you understand the real requests behind performance issues. When a metric shows an unusual spike or dip in value, an exemplar links that data point to a trace captured at the same time. This gives you immediate context for what was happening in your environment when the metric was recorded, making root cause analysis much easier.

This image shows exemplars enabled for a MAX aggregated metric, grouped by the entity_types attribute. Exemplars are indicated by a small circle icon. The sample size is the Top 100 exemplars.

OpenTelemetry supports exemplars for these types of metrics: Histograms, synchronous Sums and synchronous Gauges. To learn more about OTel-originated metrics including its type (also known as, Instrument Type), see the semantic convention for each metric. An example is http.server.request.duration. For more information on OTel-originated metrics exported by APM Libraries, see Services Metrics > Exemplars.

To view exemplars in the Metrics Explorer, search for and select the metric of interest. Then use the three-dot menu above the chart to toggle on the exemplars overlay and view them alongside your metrics chart. If accessing from Metrics tab in the Service Details view, the exemplars will be enabled by default when the metric contains exemplar annotations (otherwise, the toggle remains disabled and grayed-out).

Enabling exemplars also enables the drop down menu to change the sample of exemplars that are displayed.

  • Top N exemplars: Displays exemplars limited to N based on highest metric values.

  • Last N exemplars: Based on most recent entries.

Supported values for N are 100, 200, and 500. The default sample selected is the Top 100 exemplars.

In the example above, the sample in this image comprises the Top 100. Changing the way the data is aggregated using the tools above the chart on the upper left will update the view of the relationship between the exemplars plotted and the selected metric. Exemplars are represented by a circle icon.

This image shows the open three-dot menu with the Exemplars option toggled on. The default sample selected is the Top 100 exemplars.

Hovering over an exemplar annotation opens a context modal window that shows the timestamp and key details such as the value recorded, Service Name, and Trace ID. Here, you can open the related trace or span details to seamlessly move from aggregated metrics to detailed trace data during investigations into performance issues.

This image shows the context modal window that appears when hovering over an exemplar annotation, displaying timestamp, value, Service Name, and Trace ID.