Correlations

With Correlations, you can compare your query results to metrics data or data from a Board in the same view. It may be useful to view infrastructure metrics for your systems alongside query results from non-metrics datasets. For instance, a system running out of memory, CPU, or network resources might be the reason for an out-of-compliance SLO or an alerting trigger. Seeing the graph of the problem alongside graphs of relevant system resources could confirm or deny this kind of hypothesis. For example, when investigating a problem with your backend services, you may want to see system memory, CPU, and disk utilization in conjunction with the backend services tracing data. Use or create a Board with system memory, CPU, and disk utilization queries. Then, query your backend services tracing data, and compare it to your aforementioned Board using the Correlations tab in Query Builder.

​

Using correlations

To see metrics or board correlations:

1. Run a query.
2. Select the Correlations tab below your query results.
3. Switch between Metrics or Boards correlations using the toggle buttons in the top-right.

​

Metrics correlations

When you run an events-based query, Honeycomb automatically enables the Correlations > Metrics view beneath your query results. When you open this view, Honeycomb triggers a parallel query against your metrics dataset, using the same time window and any compatible filters from your main events query (when the fields exist in both datasets’ schemas). You get a contextual, correlated view of your metrics data shown in familiar Honeycomb visualizations. The metrics query runs independently, so you can interact with it further. Choose which metrics to access, add filters or groupings, or explore the underlying query directly.

​

Using metrics correlations

To see metrics correlations:

1. Run an events query. For example, query your user-events dataset for error rates or response times.
2. Select Correlations > Metrics below your query results.

The metrics correlations view automatically populates up to three time series charts for relevant metrics. Each chart shows a time series over the same time period as your events query. You can inspect and edit the metrics charts:

• Hover over data points to see metric values and group details.
• Adjust metric and aggregation selections using the dropdowns above each chart.
• Add filters or groupings in the Metrics view to refine what appears.
• Select the chart context menu to open the underlying metrics query in a new view.

Aggregated data points for each metric are also listed in the table below the charts.

​

Span metrics

When viewing an individual span in a trace, the Span metrics panel gives you a correlated view of infrastructure metrics and span duration data for that specific span, without leaving the trace waterfall view. This is useful for understanding whether a slow or error-prone span is related to underlying resource pressure. For example, a SELECT checkout.discounts database span taking 165ms might correlate with a spike in node CPU or memory usage on the host that served that request.

​

How span metrics work

When you select a span in the trace waterfall, Honeycomb automatically queries your metrics dataset using the same time window as the trace and any resource attributes from the span (such as k8s.node.name) that match fields in your metrics schema. The Span metrics panel appears below the trace waterfall and displays:

• Distribution of span duration: A scatter plot of all durations for spans matching the same operation, giving you a sense of whether the current span is typical or an outlier.
• Infrastructure metric charts: Up to two time series charts for correlated metrics (for example, k8s.node.cpu.usage and k8s.node.memory.usage), grouped by a resource dimension such as Node.

A Current Span marker is overlaid on each time series chart, showing exactly where in time the selected span occurred relative to the metric trend.

​

Resource type selector

Use the Resource Type dropdown in the top-right of the Span metrics panel to change the resource dimension used to group and filter the metric charts. For example, switching from Node (k8s.node.name) to another resource type updates the charts to reflect metrics by that dimension.

​

Board correlations

The Board correlations view consists of a dropdown window that displays the selected data source, a Filter search box, and up to the first six saved queries from the selected Board data source. Each query contains one or more chart visualization(s) and a bulleted list icon (). Hover over the bulleted list icon () to reveal the query’s name and composition, and an option to open the query in a new tab. Selecting anywhere in the expanded query details opens the query in a new tab.

​

Change Data Source

To change the view based on a different board, select the dropdown window and choose from a Board or a Suggested Metrics view. (Suggested Metrics views only populate if the fields in your dataset match a predetermined list.) You must create the Board beforehand in order to use it in this interface. To set a default Board view in Correlations, navigate to the Default Correlations Board setting in Dataset Settings.

​

Filter Correlations Display

Use a filter when searching for specific results. To filter, use the Select Filters… search box to enter one or more filter parameters. Suggested parameter fields appear, based on your data. Select Apply Filters to apply the parameter(s) to all the displayed queries on your board. When you apply filters, the displayed queries also update to include a WHERE clause with the filter parameter(s). For example, if you have a board called “Errors Overview”, which looks at all the errors occurring in your system, you can apply a filter to look at errors occurring in the frontend service that you own. In this case, a filter of service.name=frontend applies that parameter on top of every existing query on the board to output service-specific results. If a filter parameter does not exist as a field within a chart, then the chart displays a “No results in this time range” message. To remove a filter, use the icon () next to each filter.