Dashboard Configuration

The Dashboard is a tool for configuring observable data monitoring boards. It is designed to meet all data query requirements, ensuring that every piece of data customers need to observe can be visualized. As a monitoring board, it also requires rich visualization effects and aesthetic appeal. Therefore, the Dashboard provides a variety of charts and functional cards, allowing users to select the most appropriate ones based on their actual scenarios and desired visual outcomes.

Scenario Description

In operational scenarios, it is often necessary to display combinations of different data. In some cases, viewing data for specific objects is also required. Additionally, the Dashboard supports scenarios where customers need to compare data across time and space dimensions to observe changes.

Feature Description

Charts

Chart cards are the most commonly used components in the Dashboard. Ten different chart styles are available, each rendering based on the results returned from data queries and transformations, with varying style configuration options.

Card Styles

The Dashboard offers ten chart types: Trend, Numerical Value, List, Proportion, Ranking, Gauge, Distribution, Honeycomb, China Map, and World Map. Each card has distinct display characteristics. You can choose the appropriate card type based on your specific scenario and visualization needs. Configuration syncs across multiple cards, but note that only Trend, List, Honeycomb, and Map charts support displaying multiple data series on a single graph.

Data Query

Data querying is the foundational prerequisite for all chart card displays. The ONE platform provides two methods: Selection Query and PromQL Query.

Selection Query - Data Selection

Selection Query allows you to query data through point-and-click operations on the UI based on data models. It supports querying five data types: Metrics, Entities, Events, Logs, and Records.

The minimum query unit is a data item or expression (expressions are covered in the PromQL section). Each data item executes one query and returns results.

Metric Query Steps:

1. Select the Metric data type. In the metric model, choose a metric by navigating through Entity Category -> Entity Model -> Metric Category -> Metric Name.
2. The "My Favorites" category displays metrics marked as favorites in the metric model. Important metrics can be tagged as key metrics for quick selection.
3. When selecting a metric, its information is displayed, including metric type, baseline, unit, aggregation method, calculation formula, and description.
4. Different metric types support different aggregation methods:

   Metric Type	Aggregation Methods	Default Aggregation
   gauge	Average, Maximum, Minimum, Latest Value	Latest Value
   histogram	Average, Minimum, Maximum	Average
   sum	Sum	Sum
   count_distinct	Distinct Count	Distinct Count
   formula	Calculate	Calculate

5. After specifying a metric, querying without filters returns results for all instances. To query data under specific conditions, add filter criteria. Metric filters include three aspects:
   • Entity Dimensions: Each metric has defined related entities. Attributes of these related entities can filter the metric.
   • Metric Dimensions: Independent query dimensions specific to the metric, unrelated to entities.
   • Metric Values: Filter based on the metric value magnitude. To ensure accuracy, value filtering is applied after grouping.
6. By default, query results are aggregated into a single value. To view metrics across different dimensions, specify grouping conditions. Metric grouping includes two aspects:
   • Entity Dimensions: The downstream attributes of the metric's related entities can be used for grouping. Not all attributes are available for grouping; this depends on the storage method of the entity and metric.
   • Metric Dimensions: Independent grouping dimensions specific to the metric.

Entity/Event/Record/Log Query Steps:

1. Select the Entity/Event/Record/Log data type, then choose the object by category.
2. For these data types, both Aggregate Statistics and List Display functions are provided. List Display is only available within the List card.

Aggregate Statistics (Example: Entities):

1. Aggregate Statistics performs aggregated calculations on attributes.
2. Select the attributes to aggregate by choosing Entity Type -> Entity Attribute(s). Multiple attribute aggregations can be added within one data item. The aggregation method depends on the attribute type:
   • Text, Time: Count, Distinct Count.
   • Numerical: Average, Sum, Maximum, Minimum.
3. Filters can be applied based on entity attributes and entity tags. Operators vary by attribute value type.
4. Grouping dimensions include entity attributes and tag keys.

info

Differences between Entity/Event/Record/Log queries:

• The selection of Entity Model / Event Type / Record Type / Log Index determines the available attribute fields in the next step.

• Entity data lacks timestamp information and cannot generate trend graphs in aggregate statistics.

List Display (Example: Entities): List Display provides a quick way to build lists and is only effective in the List card style.

1. Directly select multiple attribute fields for display to quickly build a list.
2. Add filter conditions based on attribute fields.
3. List Display does not support custom grouping; it inherently groups by each detailed data item's ID, showing granular data.

Selection Query - Additional Features

Deep Analysis: Beyond displaying registered data, the Dashboard provides quick data processing methods to enhance user experience, including: common arithmetic transformations for metrics, and quick access to historical performance data.

tip

Entity/Event/Record/Log List Display shows raw data and does not support Deep Analysis.

Sorting & Limiting: The Dashboard allows selecting the data scope and quantity. Use Sorting & Limiting to sort data based on a specific metric or attribute field and display a defined number of results.

Time Aggregation:

• Trend charts automatically aggregate data based on a default granularity derived from the time picker's range. The Dashboard supports manual time granularity selection for Trend cards to meet different needs.
• List cards, which typically lack time information, can also use time aggregation to display time fields, enabling comparison of multiple metrics within the same timeframe.
• To maintain optimal data density in trend lines, the effective granularity options relate to the selected time range:

  Time Range Length	Default Granularity	Available Granularities
  (0, 3h]	1min	1min, 10min, 1h, 6h, 1d
  (3h, 1d]	10min	1min, 10min, 1h, 6h, 1d
  (1d, 7d]	1h	10min, 1h, 6h, 1d
  (7d, 14d]	6h	1h, 6h, 1d
  (14d, ∞)	1d	6h, 1d

• If the time range is too large for a fine granularity, the default granularity will be used.
• Entity data lacks timestamps and does not support time aggregation.

Expressions

Some metrics require calculation via expressions, e.g., calculating success rate from error count and total count, or computing health scores.

• Expressions support operators and functions like +, -, *, /, max, min, round, and ternary operations (which can use ||, &&, ==, >, >=, <, <= to build logical expressions). Use metric IDs and operators to build expressions.
• Special Operator Usage:
  • max: Returns the maximum value, e.g., max(a, b) returns the larger of a and b.
  • min: Returns the minimum value, e.g., min(a, b) returns the smaller of a and b.
  • round: Controls decimal places, e.g., round(1/3, 2) returns 0.33.
  • ||: Logical OR, e.g., (a>1) || (b>1) returns 1 if either condition is true, otherwise 0.
  • &&: Logical AND, e.g., (a>1) && (b>1) returns 1 only if both conditions are true, otherwise 0.
  • Ternary Operator: Returns a value based on a logical expression, e.g., a > b ? x : y returns x if a > b is true, otherwise y.
• Usage Example:
  • Add metric a1 (CPU Usage) and metric b1 (Memory Usage).
  • Expression: (a1>80 && a1<95 || b1>85 && b1<95) ? 2 : ((a1>=95 || b1>=95) ? 3 : 1)
  • Meaning: The expression evaluates to 2 if CPU usage is between 80%-95% OR Memory usage is between 85%-95%. It evaluates to 3 if CPU usage >=95% OR Memory usage >=95%. Otherwise, it evaluates to 1. This expression helps flag hosts nearing or exceeding resource thresholds.

PromQL Query

In addition to Selection Query, PromQL Query is another common method. Click PromQL Query and enter the correct query syntax to retrieve metric data.

tip

PromQL Query only supports metric data, not Entities, Events, Records, or Logs.

Data Volume Settings

• These settings apply to all queried data items, controlling query granularity via maximum data points and minimum interval.
  • Max Data Points: The maximum number of data points per series, used to calculate the step interval. Defaults to a value based on card width. For optimal display, the maximum cannot exceed 1000.
  • Min Interval: The minimum time interval between two data points. Recommended to match the data reporting frequency. Default is 15s.
  • Interval: The actual step interval used for querying, calculated as Time Range / Max Data Points.
• Example: Querying metric A for the last 1 hour in detail.
  • Default Max Data Points=1000, Min Interval=15s. Calculated Interval = 1h/1000 = 3.6s < 15s. Final Interval used = 15s.
  • If Max Data Points=50, Min Interval=15s. Calculated Interval = 1h/50 = 1.2min > 15s. Final Interval used ≈ 1min.

PromQL Syntax

The ONE platform is highly compatible with PromQL syntax (99.44%). Refer to the PromQL Syntax Guide for detailed specifications.

AI Assistant

The Dashboard incorporates AI capabilities for generating PromQL queries. You can ask the AI assistant to generate PromQL statements directly.

1. Click AI Assistant for PromQL to jump to the PromQL query interface.

2. Click AI Help within the input box, and ask the assistant in the pop-up window, e.g., "Query CPU usage for hosts over the last 5 minutes."

   

3. Click on One-Click Fill to input the generated result into the PromQL statement input box.

   

Data Transformation

The Dashboard supports transforming query results, currently offering "Merge Data Items" and "Join by Field".

Merge Data Items

This function automatically merges data items based on common keys, filling missing values with --. No configuration is needed.

Example: Data Item A:

Host	kafka broker	cpu_usage_latest
br-apm-005	10.241.20.65.9092	53.54%
br-apm-004	10.241.20.64.9092	33.79%
br-apm-003	10.241.20.63.9092	19.56%

Data Item B:

Host	OS	cpu_usage_latest
br-apm-005	windows	53.54%
br-apm-004	linux	33.79%
br-apm-002	AIX	19.56%

Result after merging:

Host	kafka broker	OS	cpu_usage_latest (A)	cpu_usage_latest (B)
br-apm-005	10.241.20.65.9092	windows	53.54%	53.54%
br-apm-004	10.241.20.64.9092	linux	33.79%	33.79%
br-apm-003	10.241.20.63.9092	--	19.56%	--
br-apm-002	--	AIX	--	19.56%

Join by Field

This function also merges data items but allows user-defined join conditions (INNER JOIN or OUTER JOIN).

Example using same Data Items A and B:

• Join Type: INNER JOIN; Condition: A.Host = B.Host; Result includes only matching rows (br-apm-005, br-apm-004).
• Join Type: OUTER JOIN; Condition: A.Host = B.Host; Result includes all rows from both tables, with -- for non-matching fields (same as "Merge Data Items" result above).

Field Display Settings

• Field Display Settings control the visibility of each field in the query results, allowing you to modify aliases and set units.
• Default Display: Shows all returned fields by default; not customizable.
• Custom Display: Manually select fields to display and configure their appearance.
  • After a query runs, all dimension fields and metric/attribute fields from the results are available for selection. Only added fields will be displayed.
  • Change the display order by dragging fields.
  • Enter custom text in the input box to set an alias for the field.
  • For numerical fields, set a unit. This replaces the original unit label without converting the value itself; the display will be scaled according to the selected unit (e.g., converting bytes to GB).

Scenarios

Scorecard

The Scorecard is a specialized card for displaying percentage-based scores (0-100), allowing you to view the health or performance score of different objects based on specific calculations.

• Predefined Scenarios: The Dashboard provides five predefined scenarios: System, Service, Native Application, Web Application, and MP Application. Selecting a scenario automatically configures the card to show the health scores for all objects within that scenario, based on rules defined in Rules Configuration > Health Assessment > Health Score.
• Generic Configuration: This mode allows custom configuration. Users can add metrics and select one data item (often an expression using ternary operators) to display as the score. Scores are capped between 0 and 100.

Topology

You can display Topology cards in the Dashboard. Select instances from categories like System, Service, Application, Database, Message Queue, or Remote Call to visualize their call relationships, consistent with the view in the Observability Insights > Global Topology module.

Trace Analysis

You can display Trace Analysis cards in the Dashboard. Select instances from categories like Service, Application, Interface, Database, Message Queue, or Remote Service Instance to visualize trace analysis graphs, consistent with the view in the System Services > Trace Analysis module.

Trace Details

You can display Trace Detail cards in the Dashboard. Enter a specific Trace ID to show the detailed view of that trace, consistent with the view in the Investigation Center > Traces > Trace Details module.

Filters

Often, users want to switch conditions dynamically across the dashboard. Filter cards enable this by providing global, interactive filters that pass conditions to other linked cards.

• Generic Filter: Select one or more fields from data models (Entity Attributes, Metric Dimensions, Event/Record/Log Attributes) as filter parameters.
• Entity Filter: Quickly select a specific entity; all attributes of that entity become available as filter parameters.
• Single-Item Filter: Similar to the Generic Filter but limited to a single field. Presents as a simple dropdown for value selection.
• Time Filter: For comparing data across different time periods, use the Time Filter to set a specific time range and link it to target cards.

After configuration, filters must be linked to other cards via Interactions. For filtering to work correctly, the target card's filterable dimensions must match the filter parameters.

Layout

To ensure an organized visual layout, the Dashboard provides three container card types: Group, Container, and Tab, in addition to query cards.

• Group: Creates an expandable/collapsible section. Collapsing the group hides all cards within it. Useful for categorizing numerous cards.
• Container: A simple container holding multiple other cards.
• Tab: Similar to a container but with multiple tabs. Clicking tabs switches between different sets of cards, ideal for organizing content by scenario.

Other

The Dashboard includes additional card types for supplementary information.

• IFrame: Embeds an external webpage via its URL.
• Text: Displays custom text, supporting variable references.
• Markdown: Renders content using Markdown syntax, supporting variable references.
• Image: Displays an image uploaded from your local machine.

---