Identify CPU and Memory Problems on a Device

Use the CPU & Memory view to compare the managed applications resource consumption to overall host resources.
Investigate device resource consumption
Note

The article describes the default dashboard view as provided by Aternity. Administrator of Aternity in your organization can completely modify the view. If your view is different from the described below, contact your Administrator of Aternity for more information.

Procedure

  1. Step 1 Open a browser and sign in to Aternity.
  2. Step 2 Use the search box in the top bar to quickly find a particular device.
    Finding a user, hostname, device, or application in Aternity

    You can enter a device hostname or IP address or enter the name of a dashboard.

    Note

    You can use spaces and other alphanumeric characters. This search is not case sensitive, so it will search for the text in any combination of upper and lower case characters. The search supports email addresses and @. The search does not support other wild cards (like *) or other forms of logic (and/or).

  3. Step 3 Select the Build Your Own tab.
    Select Build Your Own
  4. Step 4 Select the CPU & Memory tab.
  5. Step 5 Adjust the time range within which you want to troubleshoot the device.
    To do so, drag the time blue gauge or only its right or left border. Alternatively, open the time picker and manually set the timeframe.
    Select the exact time period for your analysis

    You can also use the predefined buttons: Recent Week , Recent Day , Recent Hour , or Recent 15 min . To change the timeframe of the dashboard, use the timeline gauge on the top of the screen. The arrows on the time bar shift the time and blue gauge in a week or day or else, depending on the selected preset.

  6. Step 6 Hover over the graphs to view the resources consumption at a given time.
    To see which spike is which, hover over the legend to see the colored trend line.
    View the trend for the selected attribute
  7. Step 7 Analyze the spikes by correlating data in different widgets.

    If there is no data available for a widget, change the timeframe. Probably, there are no records for the selected time. To see the data, select a wider time period.

  8. Step 8 (In the CPU Usage by Logical Core) Select a specific area on the chart and drag a mouse over it to zoom in on a certain time period. All widgets will change accordingly.

    Select Back to return to the previous time range. If you have zoomed in multiple times, select Back as many times as necessary to return to the initial view.

  9. Step 9 View all the managed applications process resources at an hourly resolution.

    The process resource KPIs are aggregated over an hour.

    The default sort is by average CPU utilization. Use the drop-down menu to define how to sort the view.

    Field Description
    CPU Util Avg

    Displays the average percent CPU usage for the device over the aggregation period, calculated from data that Aternity aggregates every two minutes.

    CPU Util Max

    Displays the utilization of the most active core during the minute.

    If there are 4 cores, one of them may utilize 100% of its CPU at a given minute and other 3 units are idle (it happens when a process gets stuck). The average CPU utilization will be 25%.

    In this case, it is very useful to troubleshoot with raw data REST API query and to check HRC_CPU_CORE_UTIL_MAX. The data also includes the ID of the core that was most active during the minute; so, it is possible to see if it is the same process gets stuck for a while (if you see the same core at 100% CPU for a few minutes, it is usually the case).

    Virtual Memory Util Avg

    (Windows only) Displays the current usage of a device's virtual memory as a percentage of the device's total virtual memory (physical RAM plus hard disk allocation for memory page faults) at a given time, calculated from data that Aternity aggregates every two minutes.

    Virtual Memory Util Max Displays the maximum Virtual Memory usage of the device within the predefined timeframe.
    Physical Memory Util Avg

    (Windows, Macs, mobile) Displays the percentage usage of the device's physical RAM memory at a given time, calculated from data that Aternity aggregates every two minutes.

    Physical Memory Util Max Displays the maximum Physical Memory usage of the device within the predefined timeframe.
    Application Minutes Running Total

    Displays the number of minutes that the process of this application has been running on this device.

    GDI Objects Total

    Displays a measure of the device's graphics resource usage in Windows. The Agent collects this value only if manually configured in your deployment.

    IO Write Rate Avg

    Displays the average rate at which the device writes to the hard disk, in kilobytes per second, calculated from data that Aternity aggregates every two minutes.

    IO Read Rate Avg

    Displays the average rate at which the device reads from the hard disk, in kilobytes per second , calculated from data that Aternity aggregates every two minutes.

    Not Responding Errors Total

    The Aternity Agent checks the application's process resource consumption (PRC) of the managed application a few times a minute to validate it's responsiveness. Then, the Agent reports the total number of times it checked the PRC (as PRC_Sampling_Count_Total value), and how many of those returned with an error (PRC_Not_Responding_Errors_Total value). This indicates there is a message loop inside a long running process which does not respond in a timely manner or completely got stuck). Divide PRC_Not_Responding_Errors_Total by PRC_Sampling_Count_Total to get an appropriate process resource unavailability metric.

    Sampling Count Total

    The Aternity Agent checks the application's process resource consumption (PRC) of the managed application a few times a minute to validate it's responsiveness. Then, the Agent reports the total number of times it checked the PRC (as PRC_Sampling_Count_Total value), and how many of those returned with an error (PRC_Not_Responding_Errors_Total value). This indicates there is a message loop inside a long running process which does not respond in a timely manner or completely got stuck). Divide PRC_Not_Responding_Errors_Total by PRC_Sampling_Count_Total to get an appropriate process resource unavailability metric.

    User Objects Total

    Displays a measure of the device's usage of resources assigned to window management. The Agent collects this value only if manually configured in your deployment.