Heat Map
Contrail Insights provides a real-time Heat Map of resources. Heat Map is a visual depiction of the relationship between hosts and instances that allows you to understand infrastructure performance at a glance.
Heat Map can also be thought of as a tool to understand usage patterns of physical infrastructure components that provide metric correlation for an ever-changing virtual infrastructure. In addition, one can consider it as a tool for visualizing the usage patterns of entities of the virtual infrastructure itself.
Using the Heat Map
Use the top context menu to select the scope of entities to display. In the following example, Infrastructure is selected, which displays all hosts. In Figure 1 there are three hosts, each represented by a rectangle. Under each host rectangle is a square for each virtual machine executing on the host. Figure 1 shows a heat map of infrastructure components displaying usage patterns.
Temperature Scale
The temperature of an entity is displayed for a metric selected from the drop-down lists for host and instance. The temperature scale is automatically determined by Contrail Insights using machine learning that evaluates values of the metric across the infrastructure.
Example: Using the Heat Map
In Figure 2, the instance.memory.usage metric is selected.
Each instance is colored according to its memory usage and the temperature scale determined by Contrail Insights. The temperature scale is displayed in metric range at the top. In Figure 2, instances are colored according to the following scale:
| Green | Using between 0-46.8% of memory capacity. |
| Yellow | Using between 46.81-70.19% of memory capacity. |
| Red | Using between 70.20-78% of memory capacity. |
The range ends at 78% in this example because that is the maximum value from the last hour across all instances. The scale changes according to the recent resource consumption learned by Contrail Insights.
Simultaneously, a host metric can be selected to show the temperature of hosts, as shown in Figure 3.
To display the exact values of metrics for an entity, place the cursor over the entity. A pop-up box displays a metric table. Scroll to the metric to view its last reported value. In Figure 4, the mouse cursor is hovering over a host to display its metric table.
Filtering Entities
Entities in the view can have multiple filters applied to them. These filters help you:
View the temperature of a subset of entities in a logical group.
Visualize how a subset of instances are distributed across hosts.
To filter resources, select a value for any given filter in the row of filters. Entities not selected by the filter will be depicted in gray. Entities that are selected by the filter will be colored according to the temperature scale.
In Figure 5, Bad is selected from the Health filter to display any entity that has bad health, according to user-defined health profiles.
The Heat Map can be further filtered to see all of the resources that belong to a specific virtual network. This can be done by selecting a virtual network from the Virtual Network filter. In Figure 6, instances attached to test-net are colored blue because an Instance Metric has not been selected.
Multiple filters can be applied at the same time. In Figure 7, instances that belong to both admin project and demo-shared-net virtual network are colored blue. All other instances are gray.
