Metrics Management¶
After completing metrics data collection, you can view all reported measurements and their associated metrics, tags, time series statistics, daily changes, and data storage policies in Metrics Management.
Measurement¶
A measurement is a collection of similar metrics, containing multiple metrics and tags. The system displays the following information for a single measurement:
Metrics¶
Metrics reflect the system's operational status (e.g., CPU usage, page load time, etc.) and consist of a metric name (identifier) and a metric value (specific numerical value).
Click the measurement name to view complete metric details (name/type/unit, etc.), and the following operations are supported:
- Filter metrics based on keywords in the search bar;
- Export the metric list as a CSV file;
- Perform one-click metric analysis in the operation column;
- Click to expand metric data and edit metric configurations.
Permissions and Priority
- Edit permissions: Only standard members and above can perform operations;
- Configuration priority: The unit and description customized in Metrics Management take precedence over the default collected data. For example, after changing the unit from
B
toGB
, the data in the scenario's chart queries will automatically be displayed inGB
.
Tags¶
The attributes of the data point collection object are identified by tags, which are presented in key-value pairs, consisting of a tag name and a tag value. A data point can have multiple tags.
Click the measurement name to view the corresponding tag list, including the tag name, tag value statistics, and descriptions, and the following operations are supported:
- Filter tags based on keywords in the search bar;
- Export the tag list as a CSV file.
Time Series¶
Time series are dynamically generated by tag combinations and include metrics, tags (fields), and data storage duration. Among them, "metric + tag (field combination)" constitutes the primary key for data storage.
In TrueWatch, the total number of time series for the current measurement is displayed here, and the unique value count, time series count, and daily change are presented through metric statistics and tag statistics charts.
Calculation Logic
The metric/tag statistics chart is calculated based on the hierarchy of overall metric → single metric → grouping of the metric by tags → specific tag values → global view of all tags + tag statistics.
In the statistics chart, the calculation logic for each field is as follows:
- Unique value count: The deduplicated count of the tag globally or within a single metric dimension (i.e., how many unique independent values the tag contains)
- Time series count: The number of time series designed for a single metric/single tag value
- 1d: The change in the number of time series compared to the previous day
- 1d%: 1d / the number of time series from the previous day
- Total proportion:
- Metric level: The time series contained by the metric + tag / the total number of time series for the metric;
- Tag level (single metric): The total number of time series for a single metric & single tag value / the total number of time series for a single metric & all tag values;
- Tag value level: The time series designed for the tag value + the metric / the total number of time series for the metric value;
- Tag level (full): The time series for the tag value / the total number of time series for the tag.
Term | Description |
---|---|
Database | A database container for storing data |
Measurement | Measurement (similar to a MySQL table) |
Field | Metric key-value pair (stores real values, not indexed) |
Field Set | A collection of multiple Fields |
Field Key | The key name of a Field (e.g., cpu_usage ) |
Field Value | The value of a Field (e.g., 85.3) |
Tag | Tag key-value pair (descriptive metadata, indexed) |
Tag Set | A collection of multiple Tags |
Tag Key | The key name of a Tag (e.g., host ) |
Tag Value | The value of a Tag (e.g., server-01 ) |
TimeStamp | Data point timestamp (nanosecond precision) |
Retention Policy | Data retention period (e.g., 30 days) |
Series | Unique time series (Measurement + Tag Set combination) |
Storage Policy¶
The storage policy is divided into two levels:
- Workspace Global Settings: Applied to all measurements
- Measurement Custom Settings: Settings for specific measurements
Modify Policy¶
The custom settings of a measurement take precedence over the global settings of the workspace. That is: After customizing the storage policy for a measurement, modifying the global settings will not affect that measurement.
Note
Each policy modification triggers immediate data deletion. Please back up important data in advance.
Considerations¶
The following key rules should be noted when modifying the storage policy of a measurement:
- Immediate data deletion: After modifying the policy, the data of the measurement under the original policy will be permanently deleted and cannot be recovered. Subsequent policy modifications will also delete the data under the current policy;
- Independent storage mechanism: After customizing the policy, the measurement data will be stored independently. Modifying the global policy duration at this time will not affect the already customized measurement;
- Restore global policy: When reverting to "global default," the data will be stored in the global library again. Subsequent modifications to the global policy will affect this measurement;
- Same duration does not mean the same policy: Even if the custom duration is the same as the global duration, the data is still stored independently. Example: Setting a measurement to 7 days when the global duration is 7 days still counts as independent storage.
Modification Path¶
Global Settings¶
Go to Management > Workspace Settings > Risky Operations > Change Data Storage Policy > Metrics, and select the storage duration.
Measurement Custom Settings¶
Go to Metrics > Metrics Management, and click the icon in the operation column of a single measurement to select the duration. If not set, it automatically inherits the Global Policy.
Important Restrictions
- Permission control: Only the workspace owner can configure;
- Data risk: Modifying the policy causes immediate permanent deletion of the data under the original policy.