Visualization types

Observe offers a rich array of visualization types for visualizing your data. Visualizations are a way to view data within dashboards and Worksheets.

Using a Kubernetes cluster observability as an example, we can see how each type of visualization can answer a specific question pertaining to our use case:

Chart or graph type	Description	Use case example
Bar chart	Good for comparing quantities across different categories.	Which namespaces are producing the most error logs today?
Change over time	Good for visualizing the change in a metric over the query window, split out by groups.	How has the total error rate changed over the last 24 hours?
Choropleth map	Good for visualizing data on a map of the world using color fills on political regions.	Which cloud regions (us-east, eu-west, ap-south) have the highest latency?
Directed acyclic graph (DAG)	Good for visualizing relationships and groupings of acyclic data.	What is the hierarchy of cluster → namespace → pod → container, and where do failures originate?
Geographic map	Good for visualizing data on a map of the world.	Where are users experiencing errors geographically?
Heatmap	A visualization that displays an aggregated value as a color (to represent a low to high gradient) within a 2D grid.	At what times of day do error spikes occur most frequently?
Hex grid	A way to visualize data grouped into sections (hexagons). Useful for getting a quick overview of a variety of objects, such as nodes or servers.	Where are requests concentrated when there are too many points to plot individually?
Histogram	A bar chart of data processed by the `histogram` verb.	What is the distribution of request latency across all services?
Line chart	Good for showing how something changes over time. Also known as time-series in other systems.	How does CPU usage for a critical workload evolve during deployment?
Pie chart	Good for visualizing Datasets that produce a complete whole, such as the number of tasks executed in a time period by a continuously running task engine.	What proportion of log volume comes from each log severity (INFO/WARN/ERROR)?
Scatter plot	Good for visualizing the relationship between two variables, or for viewing outlier data.	Do higher CPU pods also produce more error logs?
Single stat	Good for viewing the latest value for a single field.	How many pods are currently in CrashLoopBackOff?
Stacked area	Good for comparing related values in context while preventing some values from being hidden by overlapping.	How is total traffic divided across services over time?
Top list	Good for way visualizing the top groups for a specific query.	What are the top 10 pods generating the most errors right now?
Waterfall chart	Good for Datasets that have a parent/child relationship in the data, such as the Tracing/Span or OpenTelemetry/Span Datasets.	Where is latency being introduced across the request lifecycle (ingress → service → database)?

All charts and graphs in Observe can be customized by configuring option such as colors, legends, labels, and formatting. See Customize and configure your visualizations to learn how.