Containers Insight 2.0: Dashboard and Widget Guide

Overview

Containers Insight 2.0 provides a default dashboard to display a summary of cost and efficiency for all clusters. It also supports powerful customization of the dashboard. Custom widgets can be used to query rich datasets to reveal deeper insights into cost, utilization, and efficiency within Kubernetes, and OpenShift Clusters.

A dashboard appears by default. You can also create custom dashboards.

Default Dashboard

The default dashboard includes KPI widgets such as Total Cost, Idle Cost, and Efficiency Score. It also displays time series chart widgets, showing trends for the top 10 idle cost clusters, and total cost trends for the top 10 clusters. Any changes to the dashboard, such as adding or deleting widgets, updating the layout, or adjusting global filters, are automatically saved. Each user has his/ her own default dashboard that can be customized independently.

Custom Dashboard

Users can create custom dashboards by saving any existing dashboards available to them. The layout of the dashboard is structured with KPI widgets on the top, chart widgets in the middle, and a table widget at the bottom. Within each section (KPI, chart), the placement of the widgets can be adjusted. If a user deletes all dashboards, they will still receive the pre-configured built-in default dashboard. Any changes, including global filters and layout adjustments, are automatically saved.

There are several widgets that can be added to the dashboard.

Pre-Built KPI Widgets

There are four pre-built KPI widgets: Total Cost, Efficiency Score, Efficiency Score (Usage), and Idle Cost. These widgets are displayed on the default dashboard and can be added to any dashboard through the drop-down menu.

Custom KPI Widget

Users can add custom KPI widgets through the drop-down menu to add a "Custom KPI" widget.

Metric Type : Options include Total, CPU, Memory, GPU, Network TX, Network RX, FileSystem, and Persistent Volume. This shows the metrics for specific resource types.

Metric Summary : Options include Cost and Utilization. This indicates whether the KPI is based on cost metrics or utilization metrics.

Metric Allocation : Options include "Fair Share," "Allocated," "Idle," and "Efficiency." This shows whether the allocation is about fair share, allocated, idle, or efficiency of the metric type, and summary.

  • Fair Share : Provisioned resources shared proportionally based on allocation among all tenants.

  • Allocated : Resources allocated to a container; the larger of usage and request.

  • Idle : Resources not allocated and fair shared among tenants; the difference between Fair Share and Allocated.

  • Efficiency : Calculated as Allocated divided by Fair Share.

The combination of metric type, summary, and allocation determines the KPI to be created. Here are some examples:

Metric Type

Metric Summary Metric Allocation KPI

Total

Cost

Efficiency Efficiency Score = Allocated Cost / Fair Share Cost (or Total Cost)

CPU

Cost Efficiency CPU Efficiency Score = Allocated CPU Cost / Fair Share CPU Cost

CPU

Cost Fair Share CPU Fair Share Cost = Cost for the fair share CPU units

CPU

Cost Allocated CPU Allocated Cost = Cost for the allocated CPU units

CPU

Cost Idle CPU Idle Cost = CPU Fair Share Cost - CPU Allocated Cost

CPU

Utilization Efficiency CPU Utilization Efficiency = Allocated CPU Units / Fair Share CPU Units

CPU

Utilization Fair Share CPU Fair Share Units

CPU

Utilization Allocated CPU Allocated Units

CPU

Utilization Idle CPU Idle Units = CPU Fair Share Units - CPU Allocated Units

These KPIs can be filtered by clusters, namespaces, workload, workload type, container, node, or label keys. Filters are combined with "AND" operations, and only the "equals" operator is allowed.

There are also several chart widgets that can be added to your dashboard.

Pre-Built Chart Widgets

There are four pre-built chart widgets:

Efficiency for Top 10 Idle Cost Clusters

Top 10 Costs by Clusters

Top 10 Least Efficiency Score by Cluster

Total Cost Trend

Custom Chart Widgets

There are three types of custom chart widgets supported:

Custom Time Series: Displays trend information on the queried datasets, based on any combination of metric type, summary, and allocation. It can be grouped by combinations of cluster, namespace, workload, workload type, node, container, or label keys. It also allows ordering by top 10 or bottom 10.

Custom Top/Bottom 10 Widget : Displays queried datasets in a specified order, supporting only vertical bar charts.

Custom Allocated vs Idle : Displays queried datasets for allocated versus idle comparison, usable for both cost, and utilization metrics.

Customization and Layout

Users can easily customize the layout and content of their dashboards by adding, removing, and rearranging widgets. Each widget's settings and filters can be adjusted to meet specific analysis needs, providing a highly flexible and personalized visualization experience.

This guide covers the essential features and customization options available in the new Containers Insight 2.0. By utilizing these dashboards and widgets, users can gain deeper insights into their Kubernetes environments, optimizing cost, utilization, and efficiency.