Skip to content

Dashboard Performance Monitoring

This page provides insight into the runtime performance of dashboard queries. It shows how queries behave while dashboards are actively in use, helping teams identify slow, failing, or timed-out queries before they impact end users.

Dashboard Performance Monitoring captures query execution metrics based on real user activity, enabling teams to diagnose issues quickly and track performance trends over time.

This information helps teams understand how individual queries contribute to overall dashboard performance.

Why This Is Useful

Dashboards often depend on multiple queries and data sources. If one query performs poorly, it can slow down or break the entire dashboard. Without performance monitoring, these issues are typically discovered only after users report problems.

Query Performance Monitoring

Query performance monitoring is useful because it measures execution time for each dashboard query, identifies slow-running queries, surfaces failed or timed-out queries, and helps pinpoint queries that require optimization.

By providing visibility into query behavior at runtime, it allows you to address performance issues proactively instead of reactively.

Key Benefits

Using Dashboard Performance Monitoring, you can quickly identify which queries are impacting overall dashboard performance, investigate issues as they occur rather than relying on user reports, capture performance data while dashboards are actively in use, reflect real-world usage rather than test or synthetic runs, and perform live diagnostics when dashboards become slow or unresponsive.

How to Access Dashboard Performance Monitoring

In the editor, use either of the following:

  • Click Testing Page at the bottom of the Dashboard properties panel.
  • Open the menu icon in the top-left corner of the dashboard, go to Queries, and select Testing Page.

    Screenshot

The Testing Page displays performance information for dashboard queries, allowing you to monitor execution behavior while dashboards are in use.

Review query performance

The Testing Page helps you find the queries that slow a dashboard down. It has two sections: Overview and Query Results. Click the arrow beside a section heading to collapse or expand it.

Testing Page showing the Overview and Query Results sections

Overview

The Overview section summarizes the performance of all the queries on the dashboard:

Card Description
Total Queries The number of queries the dashboard runs.
Total Render Time The sum of the round-trip times of all the queries.
Average Query Time The average time a query took.
Slowest Query The name and time of the slowest query.
Queries With Errors The number of queries that returned an error. This card is highlighted in red when there are errors.

Below the cards, two tabs show each query's share of the total render time, so you can see where optimization has the most impact:

  • Heatmap: each tile is a query. The size of a tile is the query's share of the total render time, and the tile shows the query name and that percentage, for example New node 6 | 10.1%. The color of a tile shows how long the query took compared with the slowest query, using the color scale below the heatmap: blue for the fastest queries, through yellow, to green for the slowest. A query that failed is red. Hover over a tile to see the query's type, server, status, and round-trip time, and to see where it falls on the scale.

  • Pie Chart: shows two donut charts:

    • % of total render time per query: each slice is a query's share of the total render time, with the total in the center. The queries that take the most time are listed in the legend, and the remaining queries are grouped into a single Other slice.

    • Working vs errored queries: the number of queries that ran successfully compared with the number that returned an error, with the total number of queries in the center.

Switch between the tabs to see the same data from a different angle, without leaving the page.

Query Results

The Query Results grid lists each query with the following columns:

Column Description
Query Path The name of the query.
Query Type The type of query, for example query or virtual.
Connection The connection the query used.
Status Whether the query succeeded.
Error The error message, if the query failed.
Performance The round-trip time of the query, in milliseconds, with a bar that breaks the time down into stages. Hover over the bar to see the time and percentage of each stage. Queries that failed, and some that ran successfully, show no bar or time.

- Dispatch (red): the time to send the query.
- Server Duration (teal): the time the server took to run the query.
- IPC & Deserialize (yellow): the time to receive the result and convert it into a form Dashboards can use.

The footer of the grid shows totals for the columns, so you don't have to calculate them yourself. For example, it shows the number of queries, the number of distinct query types and connections, how many queries succeeded and failed, and the total time in milliseconds.

Tip

Start with the queries that have the largest share of total render time. Optimizing these gives the biggest improvement to how quickly the dashboard renders for end users.

Next Steps