Product Videos Incident Response High Cardinality Debugging

Heatmaps + BubbleUp: How They Work


Sasha Sharma [Commercial Solution Architect|Honeycomb]:

Heat maps are a form of visualization we use at Honeycomb to visualize a large number of events over different time buckets. They’re great for representing values like duration or size over time. These darker colors on your heat map represent areas of the chart where you have a larger number of events, versus the lighter colored areas where you have a smaller number of events. For example, in this chart, we can see that this dark blue area shows that most of the requests took about half a second to complete. When you roll over a heat map in Honeycomb, you can also see a distribution of the events in the time bar that appears. You can see in this histogram up top, that most requests are completed within half a second, but you can also see there is a bump to the right of that histogram that shows that there are some requests taking longer.

In a heat map, diagonal lines are rare, but they tend to mean that something about your data set is changing over time. For example, in this chart, we see that the request duration is climbing up over time. You can combine heat maps in Honeycomb with another powerful tool called BubbleUp. Simply click and drag to select an area of the heat map that is of interest to you. BubbleUp can help you understand how your selected set of data points differs from the other data points in your data set. Once you’ve made your selection, you’ll see a series of histograms appear beneath your heat map. Your selected area will be represented in orange, and the baseline events, which is the rest of your data set, will appear in blue.

Each bar chart corresponds to an attribute in your data set. For example, HTTP target, type, user ID. The order of these charts is dynamic. The attributes are stack ranked based on the disparity between your selected area and the baseline. So you’ll see charts like HTTP target and user ID appear higher in the list of these bar charts, because the disparity between your selected area and the baseline is much higher. You can also build queries using these charts. Simply click on a bar that is of interest to you, and you can add that attribute to a group by field or a filter field within the query. Thank you for watching. I hope you’ve enjoyed this video.

If you see any typos in this text or have any questions, reach out to