#Management of Monitoring Dashboards

#Function Overview

The platform provides powerful dashboard management functionality designed to replace traditional Grafana tools, offering users a more comprehensive and flexible monitoring experience. This feature aggregates various monitoring data from within the platform, presenting a unified monitoring view that significantly enhances your configuration efficiency.

#Main Features

• Supports configuring custom monitoring dashboards for both business views and platform views.
• Enables viewing publicly shared dashboards configured in platform views from business views, with data isolated based on the namespace to which the business belongs.
• Supports managing panels within the dashboard, allowing users to add, delete, modify panels, zoom in/out panels, and move panels through drag-and-drop.
• Allows setting custom variables within the dashboard for filtering query data.
• Supports configuring groups within the dashboard for managing the panels. Groups can be displayed repeatedly based on custom variables.
• Supported panel types include: trend、step line chart、bar chart、horizontal bar chart、bar gauge chart、gauge chart、table、stat chart、XY chart、pie chart、text.
• One-click import feature for Grafana dashboards.

#Advantages

• Supports user-customized monitoring scenarios without being constrained by predefined templates, truly achieving a personalized monitoring experience.
• Provides a rich array of visualization options, including line charts, bar charts, pie charts, and flexible layout and styling options.
• Integrates seamlessly with the platform's role permissions, allowing business views to define their own monitoring dashboards while ensuring data isolation.
• Deep integration with various functionalities of the container platform, enabling instant access to monitoring data for containers, networks, storage, etc., providing users with comprehensive performance observation and fault diagnosis.
• Fully compatible with Grafana dashboard JSON, allowing easy migration from Grafana for continued use.

#Use Cases

• IT Operations Management: As part of the IT operations team, you can use the monitoring dashboards to unify the display and management of various performance metrics of the container platform, such as CPU, memory, network traffic, etc. By customizing monitoring reports and alert rules, you can promptly detect and pinpoint system issues, enhancing operational efficiency.
• Application Performance Analysis: For application developers and testers, monitoring dashboards offer various rich visualization options to intuitively display application running states and resource consumption. You can customize dedicated monitoring views tailored to different application scenarios to deeply analyze application performance bottlenecks and provide a basis for optimization.
• Multi-Cluster Management: For users managing multiple container clusters, monitoring dashboards can aggregate monitoring data from disparate clusters, allowing you to grasp the overall operational state of the system at a glance.
• Fault Diagnosis: When a system issue occurs, monitoring dashboards provide you with comprehensive performance data and analytical tools to quickly pinpoint the root cause of the problem. You can swiftly view fluctuations in relevant monitoring metrics based on alert information for in-depth fault analysis.

#Prerequisites

Currently, monitoring dashboards only support viewing monitoring data collected by monitoring components installed in the platform. Therefore, you should prepare as follows before configuring a monitoring dashboard:

• Ensure that the cluster for which you want to configure the monitoring dashboard has monitoring components installed, specifically the ACP Monitor with Prometheus or ACP Monitor with VictoriaMetrics plugin.
• Ensure that the data you wish to display on the dashboard has been collected by the monitoring components.

#Relationship Between Monitoring Dashboards and Monitoring Components

• Monitoring dashboard resources are stored in the Kubernetes cluster. You can switch views between different clusters using the Cluster tab at the top.
• Monitoring dashboards depend on the monitoring components in the cluster for querying data sources. Therefore, before using monitoring dashboards, ensure that the current cluster has successfully installed monitoring components and that they are operating normally.
• The monitoring dashboard will default to requesting monitoring data from the corresponding cluster. If you install the VictoriaMetrics plugin in proxy mode in the cluster, we will request the storage cluster for you to query the corresponding data for this cluster without the need for special configuration.

#Manage Dashboards

A dashboard is a collection composed of one or more panels, organized and arranged in one or more rows to provide a clear view of relevant information. These panels can query raw data from data sources and transform it into a series of visual effects supported by the platform.

#Create a Dashboard

1. Click Create Dashboard, reference the following instructions to configure relevant parameters.

Parameter	Description
Folder	The folder where the dashboard resides; you can input or select an existing folder.
Label	Label for the monitoring dashboard; you can quickly find existing dashboards by filtering through the top labels during the switch.
Set as Main Dashboard	If enabled, this will set the current dashboard as the main dashboard upon successful creation; when re-entering the monitoring dashboard feature, the main dashboard data will be displayed by default.
Variables	Add variables when creating the dashboard to reference as metric parameters in the added panels, which can also be used as filters on the dashboard homepage.

2. After adding, click Create to finish creating the dashboard. Next, you need to add variables, add panels, and add groups to complete the overall layout design.

#Import Dashboard

The platform supports direct import of Grafana JSON to convert it into a monitoring dashboard for display.

• Currently, only Grafana JSON of version V8+ is supported; lower versions will be prohibited from being imported.
• If any panels within the imported dashboard are not within the platform's supported scope, they may be displayed as unsupported panel types, but you can modify the panel's settings to achieve normal display.
• After importing the dashboard, you can perform any management actions as usual, which will not differ from panels created in the platform.

#Add Variables

1. In the variable form area, click Add.

Query

Variables of type Query allow you to filter data based on the feature dimensions of time series. The query expression can be specified to dynamically calculate and generate query results.

Parameter	Description
Query Settings	When defining query settings, besides using PromQL to query time series, the platform also provides some common variables and functions. Reference Common Functions and Variables.
Regular Expression	By using regular expressions, you can filter out the desired values from the content returned by the variable queries. This makes each option name in the variable more expected. You can preview if the filtered values meet expectations in Variable Value Preview.
Selection Settings	- Multiple Selection: When selected from the top filters on the dashboard homepage, allows the selection of multiple options simultaneously. You need to reference this variable in the query expression of the panels to view the data corresponding to the variable value. - All: If checked, an option containing All will be enabled in the filter options to select all variable data.

Constant

Constant Variables are static variables with fixed values that remain unchanged throughout the dashboard, commonly used for storing environment identifiers, fixed thresholds, or configuration parameters that need to be referenced across multiple panels without displaying as filter options.

Parameter	Description
Constant Value	The value of the constant variable.

Custom

Custom Variables allow users to define a predefined list of static options that appear as dropdown filters on the dashboard, commonly used for manual selection of specific services, teams, or categories without requiring dynamic data queries.

Parameter	Description
Custom Settings	Enter option values separated by commas, using the format display_name : value for each option (e.g., Production : prod, Staging : stage, Development : dev), or simply list values directly if display name equals value.

Textbox

Textbox Variables are variables that allow users to enter text directly, commonly used for specifying specific values or parameters that do not require dynamic data queries.

Parameter	Description
Textbox Value	The default value of the textbox variable.

2. Click OK to add one or more variables.

#Add Panels

Add multiple panels to the currently created monitoring dashboard to display data information for different resources.

Tip: You can customize the size of a panel by clicking the lower right corner; click anywhere on the panel to rearrange the order of the panels.

1. Click Add Panel, reference the following instructions to configure relevant parameters.

• Panel Preview: The area will dynamically display the data information corresponding to the added metrics.

• Add Metric: Configure the panel title and monitoring metrics in this area.

• Adding Method: Supports using built-in metrics or using natively customized metrics. Both methods will take the union and be effective simultaneously.

  • Built-in Metrics: Select commonly used metrics and legend parameters built into the platform to display the data information under the current panel.
    • Note: All metrics added to the panel must have a unified unit; it is not possible to add metrics with multiple units to one panel.
  • Native: Customize the metric unit, metric expression, and legend parameters. The metric expression follows PromQL syntax; for details, please refer to PromQL Official Documentation.

• Legend Parameters: Control the names corresponding to the curves in the panels. Text or templates can be used:

  • Rule: The input value must be in the format {{.xxxx}}; for example, {{.hostname}} will replace it with the value corresponding to the hostname label returned by the expression.
  • Tip: If you input an incorrectly formatted legend parameter, the names corresponding to the curves in the panel will be displayed in their original format.

• Instant Switch: When the Instant switch is turned on, it will query instant values through Prometheus's Query interface and sort them, as in statistical charts and gauge charts. If off, it will use the query_range method to calculate, querying a series of data over a specific time period.

• Panel Settings: Supports selecting different panel types for visualizing metric data. Please refer to Manage Panels.

2. Click Save to complete adding the panels.

3. You can add one or more panels within the dashboard.

4. After adding the panels, you can use the following operations to ensure the display and size of the panels meet your expectations.

   • Click the lower right corner of the panel to customize its size.
   • Click anywhere on the panel to rearrange the order of the panels.
   • Click the Edit button to modify the panel settings.
   • Click the Delete button to delete the panel.
   • Click the Copy button to copy the panel.

5. After adjusting, click the Save button on the dashboard page to save your modifications.

#Add Groups

Groups are logical dividers within the dashboard that can group panels together.

1. Click the Add Panel drop-down menu > Add Group, and reference the following instructions to configure relevant parameters.

• Group: The name of the group.
• Repeat: Supports disabling repeats or selecting variables for the current panels.
  • Disable Repeat: Do not select a variable, and use the default created group.
  • Parameter Variables: Select the variables created in the current panels, and the monitoring dashboard will generate a row of identical sub-groups for each corresponding value of the variable. Sub-groups do not support modifications, deletions, or moving of the panels.

2. After adding the group, you can perform the following operations on the group to manage the panel display within the dashboard.
   • Groups can be collapsed or expanded to hide part of the content in the dashboard. Panels within collapsed groups will not send queries.
   • Move the panel into the group to allow that panel to be managed by that group. The group will manage all panels between it and the next group.
   • When a group is folded, you can also move all panels managed by that group together.
   • The folding and unfolding of groups also constitutes an adjustment to the dashboard. If you want to maintain this state when reopening this dashboard next time, please click the Save button.

#Switch Dashboards

Set the created custom monitoring dashboard as the main dashboard. When re-entering the monitoring dashboard feature, the main dashboard data will be displayed by default.

1. In the left navigation bar, click Operations Center > Monitoring > Monitoring Dashboards.

2. By default, the main monitoring dashboard is entered. Click Switch Dashboard.

3. You can find dashboards by filtering through labels or searching by name, and switch main dashboards via the Main Dashboard switch.

#Other Operations

You can click the operation button on the right side of the dashboard page to perform actions on the dashboard as needed.

Operation	Description
YMAL	Opens the actual CR resource code of the dashboard stored in the Kubernetes cluster. You can modify all content in the dashboard by editing parameters in the YAML.
Export Expression	You can export the metrics and corresponding query expressions used in the current dashboard in CSV format.
Copy	Copies the current dashboard; you can edit the panels as needed and save it as a new dashboard.
Settings	Modifies the basic information of the current dashboard, such as changing labels and adding more variables.
Delete	Deletes the current monitoring dashboard.

#Manage Panels

The platform provides various visualization methods to support different use cases. This chapter will mainly introduce these panel types, configuration options, and usage methods.

#Panel Description

No.	Panel Name	Description	Suggested Use Cases
1	Trend Chart	Displays the trend of data over time via one or more line segments.	Shows trends over time, such as changes in CPU utilization, memory usage, etc.
2	Step Line Chart	Builds on the line chart by connecting data points with horizontal and vertical segments to form a step-like structure.	Suitable for displaying the timestamps of discrete events, such as the number of alerts.
3	Bar Chart	Uses vertical rectangular bars to represent the magnitude of data, where the height of the bars represents value.	Bar charts are intuitive for comparing value differences, beneficial for discovering patterns and anomalies, suitable for scenarios focusing on value changes, such as the number of pods, number of nodes, etc.
4	Horizontal Bar Chart	Similar to the bar chart but uses horizontal rectangular bars to represent data.	When there are many data dimensions, horizontal bar charts can better utilize spatial layout and improve readability.
5	Gauge Chart	Uses half or ring shapes to represent the current value of an indicator and its proportion of the total.	Intuitively reflects the current status of key monitoring indicators, such as system CPU utilization and memory usage. It is recommended to use alert thresholds with color changes to indicate abnormal conditions.
6	Gauge Bar Chart	Uses vertical rectangular bars to display the current value of indicators and their proportion.	Intuitively reflects the current status of key indicators, such as target completion progress and system load. When multiple categories of the same indicator exist, the gauge bar chart is more recommended, such as available disk space or utilization.
7	Pie Chart	Uses sectors to display the proportional relationship of parts to the whole.	Suitable for demonstrating the composition of overall data across different dimensions, such as the proportions of 4XX, 3XX, and 2XX response codes over a period.
8	Table	Organizes data in a row-column format, making it easy to view and compare specific values.	Suitable for displaying structured multi-dimensional data, such as detailed information of nodes, detailed information of pods, etc.
9	Stat Chart	Displays the current value of a single key indicator, typically requiring textual explanation.	Suitable for showing real-time values of important monitoring indicators, such as numbers of pods, number of nodes, current alert count, etc.
10	Scatter Plot	Uses Cartesian coordinates to plot a series of data points, reflecting the correlation between two variables.	Suitable for analyzing relationships between two indicators, discovering patterns such as linear correlation and clustering through the distribution of data points, helping unearth relationships between metrics.
11	Text Card	Displays key textual information in a card format, usually containing a title and a brief description.	Suitable for presenting textual information, such as panel descriptions and troubleshooting explanations.

#Panel Configuration Description

#General Parameters

Parameter	Description
Basic Information	Select the appropriate panel type based on the selected metric data and add titles and descriptions; you can add one or more links, which can be quickly accessed by selecting the corresponding link name next to the title.
Standard Settings	Units used for native metric data. Additionally, gauge charts and gauge bars also support configuring the Total Value field, which will display as the percentage of Current Value/Total Value in the chart.
Tooltips	Tooltips are the display switch for real-time data when hovering over the panels and support selected sorting.
Threshold Parameters	Configure the threshold switch for the panels; when enabled, the threshold will be shown in selected colors in the panels, allowing for threshold sizing.
Value	Set the calculation method for values, such as the most recent value or minimal value. This configuration option is only applicable to stat charts and gauge charts.
Value Mapping	Redefine specified values, ranges,regex or special such as defining 100 as full load. This configuration option is only applicable to stat charts, tables, and gauge charts.

#Special Parameters for Panels

Panel Type	Parameter	Description
Trend Chart	Graph Style	You can choose between a line chart or an area chart as the display style; line charts focus more on reflecting the trend changes of indicators, while area charts draw more attention to changes in total and partial proportions. Choose based on your actual needs.
Gauge Chart	Gauge Chart Settings	Show Direction: When you need to view multiple metrics in a single chart, you can set whether these metrics are arranged horizontally or vertically. Unit Redefinition: You can set independent units for each metric; if not set, the platform will display units from the Standard Settings.
Stat Chart	Stat Chart Settings	Show Direction: When you need to view multiple metrics in a single chart, you can set whether these metrics are arranged horizontally or vertically. Graph Mode: You can add a graph to the stat chart to display the trend of the metric over time.
Pie Chart	Pie Chart Settings	Maximum Number of Slices: You can set this parameter to reduce the number of slices in the pie chart to lessen the interference of categories with comparatively low proportions but high quantities. Excess slices will be merged and displayed as Others. Label Display Fields: You can set the fields displayed in the pie chart labels.
Pie Chart	Graph Style	You can choose either pie or donut as the display style.
Table	Table Settings	Hide Columns: You can reduce the number of columns in the table with this parameter to focus on some primary column information. Column Alignment: You can modify the alignment of data within the column using this parameter. Display Name and Unit: You can modify the column names and units used through this parameter.
Text Card	Graph Style	Style: You can choose to edit the content you wish to display in the text card in either a rich-text editing box or HTML.

#Create Monitoring Dashboards via CLI

1. Create a new YAML configuration file named example-dashboard.yaml.

2. Add the MonitorDashboard resource to the YAML file and submit it. The following example creates a monitoring dashboard named demo-v2-dashboard1:

   kind: MonitorDashboard
   apiVersion: ait.alauda.io/v1alpha2
   metadata:
     annotations:
       cpaas.io/dashboard.version: '3'
       cpaas.io/description: '{"zh":"描述信息","en":""}' # Description field
       cpaas.io/operator: admin
     labels:
       cpaas.io/dashboard.folder: demo-v2-folder1 # Folder
       cpaas.io/dashboard.is.home.dashboard: 'False' # Is it the main dashboard?
     name: demo-v2-dashboard1 # Name
     namespace: cpaas-system # Namespace (all management view creations will occur in this ns)
   spec:
     body: # All information fields
       titleZh: 更新显示名称 # Built-in field for Chinese display name (this field is created under the Chinese language)
       title: english_display_name # Built-in field for English display name (this field is created under the English language) Built-in dashboards can set bilingual translations.
       templating: # Custom variables
         list:
           - hide: 0 # 0 means not hidden; 1 means only the label is hidden; 2 means both label and value are hidden
             label: 集群 # Built-in variable display name (label is set to the appropriate name based on the language, e.g., cluster in English)
             name: cluster # Built-in variable name (unique)
             options: # Define dropdown options; if a query retrieves data, it will use requested data; otherwise, it will use options. A default value can be set (generally only used for setting default values)
               - selected: false # Whether to default select
                 text: global
                 value: global
             type: custom # Custom variable type; currently, only built-in (custom) and query are supported (Importing Grafana will support constant custom interval (after import, it will be changed to a custom variable and will not support auto))
   
           - allValue: '' # Select all, passing options with the format xxx|xxx|xxx; can set allValue for conversion (Grafana retrieves all data for the current variable as xxx|xxx|xxx, adjustments will ensure consistency)
             current: null # Current value of the variable; if not set, defaults to the first in the list
             definition: query_result(kube_namespace_labels) # Query expression for data retrieval
             hide: 0 # 0 means not hidden; 1 means only the label is hidden; 2 means both label and value are hidden
             includeAll: true # Whether to select all
             label: ns # Built-in variable display name
             multi: true # Whether multiple selections are allowed
             name: ns # Variable name (unique)
             options: []
             query: ''
             regex: /.*namespace=\"(.*?)\".*/ # Regex expression for extracting variable values
             sort: 2 # Sorting: 1 - ascending alphabetical order; 2 - descending alphabetical order (only these two support temporarily); 3 - ascending numerical order; 4 - descending numerical order
             type: query # Custom variable type
       time: # Dashboard time
         from: now-30m # Start time
         to: now # End time
       repeat: '' # Row repeat configuration; chooses custom variable
       collapsed: 'false' # Row collapsed or expanded configuration
       description: '123' # Description (tooltip after title)
       targets: # Data sources
         - indicator: cluster.node.ready # Metric
           expr: sum (cpaas_pod_number{cluster=\"\"}>0) # PromQL expression
           instant: false # Query mode true retrieves data at a specific time
           legendFormat: '' # Legend
           range: true # Default querying range when retrieving data
           refId: 指标1 # Unique identifier for display name of data source
       gridPos: # Information on the dashboard's positional layout
         h: 8 # Height
         w: 12 # Width (width corresponds to 24 grid units)
         x: 0 # Horizontal position
         y: 0 # Vertical position
       panels: # Panel data
         title: 图表标题tab # Panel name
         type: table # Panel type; currently supports timeseries, barchart, stat, gauge, table, bargauge, row, text, pie (step chart, scatter plot, bar chart, configurable through drawStyle attribute)
         id: a2239830-492f-4d27-98f3-cb7ecb77c56f # Unique identifier
         links: # Links
           - targetBlank: true # Open in a new tab
             title: '1' # Name
             url: '1' # URL address
         transformations: # Data transformations
           - id: 'organize' # Type organize; used for sorting, rearranging order, showing fields, whether to display
             options:
               excludeByName: # Hidden fields
                 cluster_cpu_utilization: true
               indexByName: # Sort
                 cluster_cpu_utilization: 0,
                 Time: 1
               renameByName: # Rename
                 Time: ''
                 cluster_cpu_utilization: '222'
           - id: 'merge' # Merging data
             options:
         fieldConfig: # For defining panel properties and appearance
           defaults: # Default configuration
             custom: # Custom graphic attributes
               align: 'left' # Table alignment: left, center, right
               cellOptions: # Table threshold configuration
                 type: color-text # Only supports text for threshold color settings
               spanNulls: false # true connects null values; false does not connect; number == 0 connects null values according to 0
               drawStyle: line # Panel types: line, bars for bar charts, points for point charts
               fillOpacity: 20 # Exists when drawStyle is area (currently does not support configuration, area defaults to 20)
               thresholdsStyle: # Configures how to display thresholds (currently only supports line)
                 mode: line # Threshold display format (area not supported currently)
               lineInterpolation: 'stepBefore' # Step chart configuration; defaults to only supporting stepBefore (stepAfter will be supported later)
             decimals: 3 # Decimal points
             min: 0 # Minimum value (currently not supported for page configuration, only supports imports that have been adapted)
             max: 1 # Maximum value (page configuration only applies to stat gauge barGauge pie)
             unit: '%' # Unit
             mappings: # Value mapping configuration (currently only supports value and range types; special types supported on data)
               - options: # Value mapping rules
                   '1': # Corresponding value
                     index: 0
                     text: 'Running' # Displayed as Running when value is 1
                 type: value # Value mapping type
               - options: # Range mapping rules
                   from: 2 # Range start value
                   to: 3 # Range end value
                   result: # Mapping result
                     index: 1
                     text: 'Error' # Values from 2 to 3 will display as Error
                 type: range # Mapping type for range
               - type: special # Mapping type for special scenarios
                 options:
                   match: null # nan null null+nan empty true false
                   result:
                     text: xxx
                     index: 2
             thresholds: # Threshold configuration
               mode: absolute # Threshold configuration mode, absolute value mode (currently only supports absolute and percentage mode; percentage mode is not supported yet)
               steps: # Threshold steps
                 - color: '#a7772f' # Threshold color
                   value: '2' # Threshold value
                 - color: '#007AF5' # Default value with no value is the Base
           overrides: # Override configuration
             - matcher:
                 id: byName # Match based on field name
                 options: node # Corresponding name
               properties: # Override configuration; id currently only supports displayName unit
                 - id: displayName # Display name override
                   value: '1' # Overridden display name
                 - id: unit # Unit override
                   value: GB/s # Unit value
                 - id: noValue # No value display
                   value: No value display
         options:
           orientation: horizontal # Control the layout direction of panels; applies to gauge and barGauge (stat will be supported later)
           legend: # Legend configuration
             calcs: # Calculating methods (only displays when the legend position is on the right)
               - latest # Currently only supports most recent value
             placement: right # Legend position (right or bottom; defaults to bottom)
             placementRightTop: '' # Configuration for the upper right
             showLegend: true # Whether to display the legend
           tooltip: # Tooltips
             mode: multi # Mode dual selection (only multi-mode supported) All data displayed when the mouse hovers over
             sort: asc # Sorting: asc or desc
           reduceOptions: # Value calculating method (used for aggregating data)
             calcs: # Calculating methods (latest, minimum, maximum, average, sum)
               - latest
             limit: 3 # Pie limits the number of slices
           textMode: 'value' # Stat configuration; defines style for displaying metric value; options are auto, value, value_and_name, name, none (currently not supported in the page configuration, but supported in imports)
           colorMode: 'value' # Stat configuration; defines color mode for displaying metric values; options are none, value, background (defaults to value; not supported in configuration but adapted in import)
           displayLabels: ['name', 'value', 'percent'] # Fields displayed in pie chart labels
           pieType: 'pie' # Pie chart type; options are pie and donut
           mode: 'html' # Text chart type mode; options are html and richText
           content: '<div>xxx</div>' # Content for text chart type
           footer:
             enablePagination: true # Table pagination enabled

#Common Functions and Variables

#Common Functions

When defining query settings, besides using PromQL to set queries, the platform provides some common functions as follows for your reference in customizing query settings.

Function	Purpose
label_names()	Returns all labels in Prometheus, e.g., label_names().
label_values(label)	Returns all selectable values for the label name in all monitored metrics in Prometheus, e.g., label_values(job).
label_values(metric, label)	Returns all selectable values for the label name in the specified metric in Prometheus, e.g., label_values(up, job).
metrics(metric)	Returns all metric names that satisfy the defined regex pattern in the metric field, e.g., metrics(cpaas_active).
query_result(query)	Returns the query result for the specified Prometheus query, e.g., query_result(up).

#Common Variables

While defining query settings, you can combine common functions into variables to quickly define custom variables. Here are some common variable definitions available for your reference:

Variable Name	Query Function	Regular Expression
cluster	label_values(cpaas_cluster_info,cluster)	-
node	label_values(node_load1, instance)	/(.*?):.*/
namespace	query_result(kube_namespace_labels)	/.*namespace=\"(.*?)\".*/
deployment	label_values(kube_deployment_spec_replicas{namespace="$namespace"}, deployment)	-
daemonset	label_values(kube_daemonset_status_number_ready{namespace="$namespace"}, daemonset)	-
statefulset	label_values(kube_statefulset_replicas{namespace="$namespace"}, statefulset)	-
pod	label_values(kube_pod_info{namespace=~"$namespace"}, pod)	-
vmcluster	label_values(up, vmcluster)	-
daemonset	label_values(kube_daemonset_status_number_ready{namespace="$namespace"}, daemonset)	-

#Variable Use Case One

Using the query_result(query) function to query the value: node_load5, and extract the IP.

1. In Query Settings, fill in query_result(node_load5).

2. In the Variable Value Preview area, the preview example is node_load5{container="node-exporter",endpoint="metrics",host_ip="192.168.178.182",instance="192.168.178.182:9100"}.

3. In Regular Expression, fill in /.*instance="(.*?):.*/ to filter the value.

4. In the Variable Value Preview area, the preview example is 192.168.176.163.

#Variable Use Case Two

1. Add the first variable: namespace, using the query_result(query) function to query the value: kube_namespace_labels, and extract the namespace.

   • Query Settings: query_result(kube_namespace_labels).

   • Variable Value Preview: kube_namespace_labels{container="exporter-kube-state", endpoint="kube-state-metrics", instance="12.3.188.121:8080", job="kube-state", label_cpaas_io_project="cpaas-system", namespace="cert-manager", pod="kube-prometheus-exporter-kube-state-55bb6bc67f-lpgtx", project="cpaas-system", service="kube-prometheus-exporter-kube-state"}.

   • Regular Expression: /.+namespace=\"(.*?)\".*/.

   • In the Variable Value Preview area, the preview example includes multiple namespaces such as argocd, cpaas-system, and more.

2. Add the second variable: deployment, and reference the variable created earlier:

   • Query Settings: kube_deployment_spec_replicas{namespace=~"$namespace"}.

   • Regular Expression: /.+deployment="(.*?)",.*/.

3. Add a panel to the current dashboard and reference the previously added variables, for example:

   • Metric Name: pod Memory Usage under Compute Components.

   • Key-Value Pair: kind: Deployment, name: $deployment, namespace: $namespace.

4. Once you have added the panels and saved them, you can view the corresponding panel information on the dashboard homepage.

#Notes When Using Built-in Metrics

• workload.cpu.utilization

• workload.memory.utilization

• workload.network.receive.bytes.rate

• workload.network.transmit.bytes.rate

• workload.gpu.utilization

• workload.gpu.memory.utilization

• workload.vgpu.utilization

• workload.vgpu.memory.utilization