Because dashboard layouts can contain different widgets (and steps), run the inspector on each layout. To open the dashboard inspector, preview the dashboard while browsing the app, select the right layout, and then click Dashboard Inspector.
- Query Metrics
- Use the query metrics to determine which step queries are taking the longest time to run. If needed, tune the step queries to increase dashboard performance.
- The inspector shows the following metrics for every step included in the dashboard
Metric Description # Runs Number of times Wave executed the step query. First (MS) Number of milliseconds that Wave took to execute the step query for the first time. Avg (MS) Average number of milliseconds that Wave took to execute the query.
- For tips about tuning SAQL queries, see Query Performance.
- Step Details
- To view the details, select a step in the inspector. The step details summarize different parts of the query, like measures, filters, groupings, and the dataset on which its built. Use the metrics in the Last Run Metrics section to identify other events that contribute to the total time required to populate the results of a step.
Metric Description Metadata Number of milliseconds that Wave took to retrieve the extended metadata. Waiting Number of milliseconds that Wave took to resolve bindings. Query Number of milliseconds that Wave took to run the query.
- Executed Query
- Wave modifies step queries based on interactions with other widgets. If you have
questions about why a step doesn’t show all expected results or behaves in an
unexpected manner, click .
- The Query Details page allows you to compare the query defined in your dashboard against the actual query executed when the dashboard is viewed. You can see what additional filters are applied due to faceting and global filters, and the resolved values of bindings, if applicable. If bindings are used in the query, the bindings tracer provides a step-by-step breakdown of each bindings function and the resolved value. The bindings tracer allows you to see if the resulting data—including both content and format—is what you intended.