Metric

Metric, also known as index, is a standard to measure the status or condition of the system in an aspect, such as the performance of the system, service or program. Metrics generates a series of data points as time goes on. The BCM provides a large number of metrics by default, such as BCC and RDS. Refer to the Cloud Product Metric for details. The users can also define metrics by themselves, and push the data of the API or bcm-tool with BCM to BCM.

• Metric name: Metric name is the string type. Metric name, scope and dimension are combined to uniquely determine one sequence of metric data.
• Metric data: Value type, the current storage period of Metric data is 40 days. The BCM acquires the data of all preset metrics by default, and the users can also push the data to the BCM monitoring platform by the API or bcm-tool.

Scope

The scope refers to the visual range of metrics. Different metrics of scopes are completely isolated from each other to avoid naming conflicts. Generally, the metrics of different applications are placed in different scopes. The scope must be designated where metrics are used.

• Preset Scope: For different cloud product monitoring data corresponding to the scope with the prefix of BCE_, refer to Cloud Product metrics and Parameters for details.
• Custom Scope: The users can create the scope by themselves. The name of custom scope cannot start with "BCE" or "BCM", and can only push data to the custom scope, not to the preset scope by the system. Refer to the Create Custom Monitor for details.

Dimension

The Dimension is used to describe various key value pairs (Key/Value, abbreviated as K/V pair) of metric characteristics. For example, the dimension of BCC metric - CPUUsagePercent is BCC instance ID (InstanceId). When determining the instance ID, we can acquire the monitoring data of CPU usage rate of the corresponding BCC instance.

The Dimension is a structure, and its definition is as below:

Data Point

The data point is the value of metric at a certain time point or during a certain period, with the format as below:

Time Stamp

For the definition of time stamp, refer to the General Description - Date and Time Specification. Each data point of the metric must have a time stamp, and use the UTC time, e.g.: 2014-09-03T08:08:08Z. The users can only push the data within the past 2 weeks (depending on the saving time of the system data) and 2 hours in the future.

Period

The Period, also known as time quantum, is used as the time span for data statistics. Generally, all the data in the period are computed according to a certain statistical function. The time quantum is measured in seconds, and the minimum accuracy is 1 minute. So it must be an integral multiple of 60 within the designated time quantum. The minimum optional value is one minute (60), and the maximum optional is 2 weeks (1,209,600). The default value is 60.

The Period is mainly used to:

• In the data query, obtain the aggregated data with different granularities within the designated time quantum.
• In setting of the alarm rules, obtain the aggregated data with different granularities within the designated time quantum, and then make corresponding comparisons with the threshold.

Statistic

The statistic refers to the aggregation of metric data within the designated time quantum . For example, one machine aggregates the CPU usage rate data by hour.

Currently, the provided statistical methods include (using the naming rule of lowerCamel):

• minimum: The minimum value of metric data within the designated time quantum.
• maximum: The maximum value of metric data within the designated time quantum.
• sum: The sum value of metric data within the designated time quantum.
• average: The average value of metric data within the designated time quantum, with the computation method of sum/sampleCount.
• sampleCount: The sample number of metric data within the designated time quantum.