Traditional object management webpages typically use a list-based format, as exemplified by the resource object list for TKE clusters in the TKE console. However, this approach suffers from several limitations, including low readability due to abstract numerical values, a lack of clear object identification, and the inability to support certain sorting methods. To break these limits, TKE has introduced Workload Map, a cloud-native asset management platform that displays all resource objects of users in a visualized way. This platform provides rich filtering, aggregation, and status display features that enable users to rapidly locate their required objects.
Workload Map displays the status and metrics of workloads by using visualized charts and diagrams on webpage. This assists users in comprehending the configuration and usage of their current workloads, as well as in identifying any potential issues.
Overview
Workload Map does not offer monitoring data for recommended Pod values on super nodes. When Workload Map aggregates monitoring metrics, if any Pod in a workload resides on a super node, statistical data for recommended values at the workload and cluster levels will be incomplete.
2. Choose Cloud native asset management > Workload Map in the left sidebar.
3. On the Workload Map page, select Region, Cluster type and Cluster options as needed.
Workload Map Features
The Workload Map page is divided into two parts: Overview and Resource Object Heat Map.
The upper part provides an overview of the workloads in the selected cluster.
1. The lower part displays individual workload resource objects in the current cluster.
Workload overview
Workload Overview: Displays the number and percentage of workloads in the current cluster that have recommended request values, as well as the total number of CPU cores and memory in nodes in the current cluster.
Note
You need to enable the Request Recommendation feature in order to receive recommended request values for adjusting workloads.
CPU Usage: Shows the CPU utilization of workloads in the cluster.
Request: The sum of CPU requests for all Pods in all workloads in the cluster.
Usage: The sum of actual CPU utilization for all Pods in all workloads in the cluster.
Recommendation: The sum of recommended CPU utilization for all Pods in all workloads in the cluster. To receive this value, Request Recommendation needs to be enabled beforehand.
Memory Usage: Shows the memory usage of workloads in the cluster.
Request: The sum of memory requests for all Pods in all workloads in the cluster.
Usage: The sum of actual memory usage for all Pods in all workloads in the cluster.
Recommendation: The sum of recommended memory usage for all Pods in all workloads in the cluster. To receive this value, Request Recommendation needs to be enabled beforehand.
Viewing workloads
You can filter workloads by metric or by status and aggregate workloads.
Filter by metric: You can filter workloads by relevant metrics. If you specify no metrics, all workloads are selected by default. You can filter workloads by multiple metrics, and the intersection of those filters is selected.
Aggregate: The filtered workloads are grouped. Workloads in the same group are shown in the same light-colored box and have the same metric value.
Status: You can filter workloads based on their status, such as the peak CPU utilization, average CPU utilization, and CPU packing density.
Peak CPU utilization: Shows the peak CPU utilization of a workload, with metrics available for the last 24 hours, 7 days, and 30 days.
Average CPU utilization: Shows the average CPU utilization of a node, with metrics available for the last 24 hours, 7 days, and 30 days.
CPU packing density: Shows the average CPU packing density of a node, with metrics available for the last 24 hours, 7 days, and 30 days.
Note
In the Workload Object Heat Map, nodes are sorted in ascending order by default, based on the currently selected "status" attribute value. For instance, if you choose "24-hour average CPU utilization" as the status attribute, workloads are displayed in ascending order based on their 24-hour average CPU utilization.
Filtered workloads are distinguished by three colors: green, blue, and red. To adjust the threshold ranges for these three display colors, you can click
in the bottom-right corner of the page.
Download: You can click the
icon on the right of Workload Overview to download the workload information list on this page.
Description of fields in the list:
Name: Name of the workload.
Namespace: Namespace to which the workload belongs.
Workload type: Kubernetes type to which the workload belongs.
24-hour average CPU utilization: Average CPU utilization (%) of the workload in the last 24 hours.
7-day average CPU utilization: Average CPU utilization (%) of the workload in the last 7 days.
30-day average CPU utilization: Average CPU utilization (%) of the workload in the last 30 days.
24-hour peak CPU utilization: Peak CPU utilization (%) of the workload in the last 24 hours.
7-day peak CPU utilization: Peak CPU utilization (%) of the workload in the last 7 days.
30-day peak CPU utilization: Peak CPU utilization (%) of the workload in the last 30 days.
24-hour average memory usage: Average memory usage (%) of the workload in the last 24 hours.
7-day average memory usage: Average memory usage (%) of the workload in the last 7 days.
30-day average memory usage: Average memory usage (%) of the workload in the last 30 days.
24-hour peak memory usage: Peak memory usage (%) of the workload in the last 24 hours.
7-day peak memory usage: Peak memory usage (%) of the workload in the last 7 days.
30-day peak memory usage: Peak memory usage (%) of the workload in the last 30 days.
Workload operations
By hovering over a workload, you can view the details of the current workload and perform the following operations:
1. Click Details to view the details about the current workload.
Workload details
Pod details
Feature switches
Request Recommendation: Cluster-level feature switch. If the switch is turned on, suitable request values are recommended for workloads based on their resource usage.
**Recommended Value (in blue font)**: Indicates the recommended request value for the current workload. You can update this value by clicking the switch. The value is displayed only when Request Recommendation is enabled and there is a recommended value.
Workload information
24-hour Average Utilization: Average CPU and memory utilization of the workload in the last 24 hours.
24-hour Peak Utilization: Peak CPU and memory utilization of the workload in the last 24 hours.
Workload chart information
CPU Usage: CPU utilization of the current workload.
Request: Sum of requested CPU for all Pods in the current workload.
Usage: Sum of actual CPU utilization for all Pods in the current workload.
Recommendation: Sum of recommended CPU utilization for all Pods in the current workload. To receive this value, Request Recommendation needs to be enabled beforehand.
Memory Usage: Memory usage of the current workload.
Request: Sum of requested memory for all Pods in the current workload.
Usage: Sum of actual memory usage for all Pods in the current workload.
Recommendation: Sum of recommended memory usage for all Pods in the current workload. To receive this value, Request Recommendation needs to be enabled beforehand.
CPU usage chart
Request: Number of requested Pods.
Usage: Number of Pods actually used.
Recommendation: Recommended usage of Pods.
Pod details
Pod name: Name of the Pod.
Pod status: Status of the Pod.
Node: Node to which the current Pod belongs.
Node type: Type of the node to which the current Pod belongs.
2. Click Recommend to show the recommended request value for the current workload. This button is available only when Request Recommendation is enabled and there is a recommended value.
Yes
No
Was this page helpful?