The agent and gateway monitor resource metric data, which can raise an alert when metric value exceeds a specified threshold. After you create a metric, create a monitor of the same collector type.
To create a metric, configure the following parameters on the CREATE METRIC page:
Attributes | Description |
---|---|
Metric Scope | Partner- or client-Specific metric. |
Adapter Type | Example: VMware, Cisco UCS, and XenServer. |
Name | Unique name associated with the metric. |
Tag Name | User-defined name for filtering. |
Display Name | Unique metric display name. |
Description | Metric description. |
Data Point type | Metric data type. |
Category | Select the appropriate category for the metric. |
Units | Metric value unit. |
Unit Multiplication Factor | Value to multiply the metric by based on the specified **Units**. |
Data Point Conversion | Metric data conversion mechanism:
|
Metric Processing | Metric processing specification:
|
Metric name format
The following are acceptable formats for a metric name:
- Metric names should not contain a dot (.), instead use an underscore (_) between the words.
system.cpu.usage
is not validsystem_cpu_usage
is valid - A metric name can have a maximum of ten underscores “_”. A metric name cannot start or end with an underscore.
- A metric name can contain numeric values (0-9), but cannot start with a numeric value.
- A metric name should be case-insensitive.
- The length of the metric name should be less than or equal to 255, including underscores.
Graphs with New Metrics data
Property | Expected Behaviour | Workaround / Details |
---|---|---|
Graphs plotting frequency Vs Monitoring Frequency | The graph shows raw samples based on default step value: 1min i..e the plotting frequency for graphs irrespective of configured monitoring frequency. |
|
Graphs plotting behavior |
| To avoid plotting gaps, ensure that the metric collection frequency matches the default graphing frequency or adjust the graphing frequency to match the metric collection intervals which is less than or equal to 5 minutes. The recommend monitoring frequency for any metric should be <= 5 minutes |
Visualize Real-time data | The data points for new metric graph data are calculatedly plotted every 1 min except at the frequency time-interval. In order to view real-time data, the metric graphs offer displaying the actual timestamps when the data was collected.Status and availability metrics need to show when data collection stopped, indicating a resource or service outage. | Graph > Expand Info section > shows “Real Time Data” option. |
Metrics Details | In the new metrics system, templates that describe the relationship between different metrics are not available. User might see the legacy metrics name while detailing, which is expected behavior and these will be slowly transitioned to show in the new metric nomenclature. | To view the relationships between template and metrics, navigate to the Monitors tab where detailed information is provided. |
Default Time Period for Graphs | Default time-period has been set as 1 Hour from the existing 4 Hr window | Default Time-period is now set 1H |
Multiple metrics | The new metric visualization supports only a single metric graph data. | The new metric visualization doesn’t support multiple metric graph data visualization due pagination and hence not available. |