Monitoring service performance
If a cloud administrator has enabled monitoring for a service (as a default behavior for the service or as an option you select as part of your service instance), you can view charts from the Performance page to see detailed information about the health of your provisioned servers of that service.
This topic discusses:
The following video (3:59) describes how you can monitor service health (or performance) in BMC Cloud Lifecycle Management.
Use the Show field to show aggregate data for different tiers in the service.
Types of charts
For VMware services, the System CPU Utilization and System Memory Utilization line charts are displayed by default.
For Amazon services, the System CPU Utilization line chart is displayed by default.
You can add charts to monitor other aspects of a service. Charts you select for a service are applied to all of your provisioned instances of that service. Charts selected by a different user apply to that user's provisioned instances of the service. The full list of available chart types includes:
|Chart type||Available metrics||Notes|
|Disk Read throughput||Average||For Amazon Web Services, only ephemeral storage is supported.|
|Disk Reads||Average||For Amazon Web Services, only ephemeral storage is supported.|
|Disk Write throughput||Average||For Amazon Web Services, only ephemeral storage is supported.|
|Disk Writes||Average||Not applicable|
|Network Received||Average||Not applicable|
|Network Transmitted||Average||Not applicable|
|System CPU Utilization||Average, Minimum, and Maximum||The Minimum and Maximum metrics are available only when you are viewing a chart for a particular tier.|
|System Memory Utilization||Average, Minimum, and Maximum|
This chart is not available for Amazon Web Services.
The Minimum and Maximum metrics are available only when you are viewing a chart for a particular tier.
You can configure the way a chart appears and performs by using the icons available in the chart, as detailed in the following table:
|Click Zoom In to open a larger view of a chart.|
|Click Refresh to update the chart with the most recently available data.|
Click Live to display data as it arrives according to the interval that the cloud administrator defined for this service. This feature is best used for critical servers with frequent polling intervals. Click Live again to stop the live stream of data.
This button is disabled for charts configured to show only historical data.
|Click Edit to change the amount of data displayed and the polling interval for the chart.|
|Click Delete to remove the chart from your display.|
How data is collected for charts
When you add or edit a chart, you can determine these key aspects for the data that is displayed:
- A specified time period for the data (scope of the chart)
- The polling interval for retrieving data (number of data points used to populate the chart)
A more frequent interval can provide a more granular view of the performance of a service. Note that server performance data is collected at a minimum interval defined by the cloud administrator as part of a monitoring level. The polling intervals that are available to select are multiples of the monitoring-level interval that the cloud administrator defined.
The Average, Min, and Max measurements in your charts represent the average, minimum, and maximum values available in the cache of data. For example, if your monitoring level has a 5-minute interval defined, the Average value in your chart shows data points that are the average of all of the data available within that 5-minute interval in the cache.
Adding a chart to a service
To reduce the impact on system performance and to streamline the chart display, select only those charts that are relevant to the data you need to monitor. For example, consider the following scenarios:
- If you want to monitor the operational state of a system on an ongoing basis, you might use the default CPU Utilization and Memory Utilization charts. If you have an application that has a high occurrence of IO, you might add a Disk Read Throughput chart. Because historical data is more relevant for the overall operational state than live data, you might configure charts to show data collected over the previous day, and at longer polling intervals.
- If you want to track data for the purpose of policy-driven actions, select a polling interval that matches the threshold selected in your monitoring rule. If you create these types of charts, consider using them on a short-term basis to call your attention to recurring performance issues. After the issues are resolved, you can remove the chart.
- You might need a chart for policy-driven activity and another chart showing the same type of data for monitoring the operational state of a system. For example, while you might use the default CPU Utilization chart for monitoring the operational state on an ongoing basis, and a second CPU Utilization chart that is set to poll for data every 15 minutes, matching the threshold value of 15 minutes set in a corresponding monitoring rule.
- You could use charts for temporary troubleshooting purposes. For example, you might create multiple Memory Utilization charts, each with a different granularity of data, to have a better view of a server's memory utilization performance over time. When the issue has been resolved, you can remove the extra charts.
To add a chart to a service
- In the Performance section, click Add Chart.
- Select a metric for the chart you want to add.
In the View data for past field, select the time interval.
For example, if you select 1 day, the chart is populated from the most recent 24-hour collection of data.
If you select 1 hour, the Select interval for data points field is set to Actual Value, the interval for which the cloud administrator predetermines.
- In the Select interval for data points field, select a polling interval.
- Click OK.
The chart is added to the Performance section.
Viewing granular data in a chart
In a chart, click the Zoom In icon .
A large version of the chart opens.
- Click and drag along the timeline at the bottom of the chart to select a time interval on which to zoom.
A gray box highlights the time period.
- Release the mouse button to zoom.
The large chart shows data only for the time you selected, with individual data points plotted on the chart. If a data point has exceeded a policy threshold, the data point appears in red.
- If necessary, repeat steps 2 and 3 to zoom again.
- To view more granular data for a particular data point, hover the mouse pointer over a data point to display a gauge with the date, time, and exact value.
- To return to the original display, click once in the timeline.
- If you were viewing a large chart, click the X icon to return to the Performance page.
Editing a chart
- In a chart, click the Edit icon
The Edit Chart dialog box opens. .
- Select values for the View data for past and Select interval for data points fields.
- Click OK.
The chart is updated on the Performance page.