Custom dashboard creation is supported in this release, along with various custom widget types. Users can also delete and add custom widgets in the default dashboard and set them as their custom dashboards.
The Treemap View provides a hierarchical visualization for clusters, enabling deeper analysis down to the container level:
- Hierarchical Visualization: Allows drilling down from clusters to namespaces, workloads and containers.
- Efficiency Score Visualization: Integrates efficiency scores within the Treemap view for quick assessments.
Both Dashboards and the Treemap View support global filters for date, clusters, and cost basis. Widgets are organized with KPIs at the top and tables at the bottom for optimal visibility.
Detailed Drill-Down Capability
Our enhanced drill-down functionality allows seamless navigation:
- From the dashboard, navigate to the table widget and select a cluster to view its namespaces in the Treemap View.
- Select a namespace in the Treemap to explore constituent workloads.
- Clicking on a workload reveals the containers within that workload.
At each level of granularity, the user can understand the different cost components and efficiency score.
For in-depth container details within a workload:
- Click on a container to view the nodes it is operating on.
- .
New Cost Column: Miscellaneous Cost
As part of this release, we are introducing a new column on the Container Insights page called the 'Miscellaneous Cost' column, available in the Table widget. This column will reflect cluster-specific costs and will initially be null for all CSPs. We are working on making this first available for GCP, with plans to expand support to other cloud service providers (CSPs) in the near future. These costs encompass additional expenses associated with containerized applications that extend beyond nodes, volumes, and data transfer. They include various services and resources necessary for cluster operations, providing a comprehensive view of all costs related to running containerized applications.
The Cost Efficiency Score is a key performance indicator (KPI) that measures the overall allocation within a containerized environment. It compares the utilized cost against what is deemed the fair share or total cost. Understanding this score involves examining three main types of costs: fair share cost, utilization cost, and idle cost.
Fair Share Cost / Total Cost
Fair share cost represents the proportion of the total node cost attributed to a container based on its allocated percentage of each resource (CPU, Memory, GPU). This is calculated by multiplying the node cost for each resource by the fair share percentage of that resource allocated to the container. The total fair share cost for a container is the sum of these amounts for all resources.
Utilization cost reflects the actual cost of resources used by a container. It is calculated by multiplying the node cost of each resource by the utilization percentage of that resource on the node. Like fair share cost, the total utilization cost for a container is the aggregate of utilization costs for all resources.
Idle cost is derived by subtracting the utilization cost from the fair share cost. It represents the cost of resources that are allocated but not utilized, indicating inefficiency.
Cost Efficiency Score Calculation
The Cost Efficiency Score is calculated by comparing utilized cost to fair share cost, encapsulating the efficiency of resource usage across all resources within a container, namespace, workload, or cluster. This score helps identify inefficiencies and potential areas for optimization, such as cluster or node rightsizing or improving workload pod affinity constraints.
A lower Cost Efficiency Score indicates greater inefficiency, suggesting more significant opportunities for cost savings and optimization within the environment. Please reach out to support if you have questions around this feature or have questions.