Manual Chapter : Analyzing Statistics Data

Applies To:

Show Versions Show Versions

BIG-IQ Centralized Management

  • 7.0.0
Manual Chapter

Analyzing Statistics Data

Browse through your managed devices looking for high resource usage

Before you can analyze BIG-IP device performance data:
  • There must be a BIG-IQ data collection device configured for the BIG-IQ system that manages your BIG-IP devices.
  • You must have discovered the BIG-IP devices that you want to analyze, and statistics collection must be enabled for those devices.
  • It is also a good idea, though not a requirement, to define the retention policy for the statistics data you are collecting.
You can use the Device dashboards to review the resource usage and traffic management status for your BIG-IP devices to examine any changes that can impact your BIG-IP services. In this example, we examine how to identify a device with high CPU usage, and its impact on traffic management. You can use this procedure to identify changes in resources such as, memory, disk activity, disk usage and interface activity.
  1. At the top of the screen, click
    Monitoring
    .
  2. On the left, click
    DASHBOARDS
    Device
    Health
    .
    The Device Health overview screen opens.
  3. In the time controls, click the
    Events
    button to display event information in the charts.
  4. On the Device Health overview screen, adjust the dimensions pane so that it fills up at least half of the screen.
  5. In the dimensions pane, click the down arrow on the BIG-IP Host Names dimension to expand the list of BIG-IP devices.
  6. In the CPU column click the
    System
    heading to sort the list of BIG-IP devices by CPU usage.
  7. If you find a device that is consuming a high level of CPU cycles, select it.
    The chart pane displays statistics for only the selected device.
  8. Scan the CPU chart for the distressed device, and look for the point where the cycles jumped up, or events that may have triggered changes in CPU usage.
    If you find a spike in usage, you can narrow the time focus. This allows you to focus on specific events, if multiple events occurred. There are two ways you can zoom in on an area of interest:
    • Click either end of the focus element on the time selector to define the interval you want to examine.
    • Click and drag over the interval in the chart itself. A small magnifying glass pops up over the highlighted area. Click the magnifying glass set the time focus to the area you highlighted.
    When you zoom in to focus on the area of interest, the focus changes for all charts on the Device Health overview screen.
  9. Make a note of both the host name of the BIG-IP device, and the time that the CPU cycles climbed.
    You have found a device that is in distress and identified the time that the issue started.
  10. On the left, click
    DASHBOARDS
    Device
    Traffic
    .
    The Device Traffic overview screen opens.
  11. On the Device Traffic overview screen, adjust the dimensions pane so that it fills up at least half of the screen.
  12. In the dimensions pane, click the down arrow on the BIG-IP Host Names dimension to expand the list of BIG-IP devices.
  13. In the Name column, click the magnifying glass icon ( ), type the host name of the BIG-IP device noted in step 8, then click the magnifying glass icon again.
    The chart pane displays statistics for only the selected device.
  14. Use the time selector controls to focus in on the time noted in step 8, and look through the traffic charts and events to see if there was a spike in traffic that corresponds to the spike in CPU cycles.
    If you find that spike, you have found the reason for the spike in CPU cycles. Traffic on that device is very high. You can further identify the application traffic activity of the associated pools & pool members, and virtual servers.
  15. On the left, click
    DASHBOARDS
    Local Traffic
    Virtual Servers
    .
    The Virtual Servers overview screen opens.
  16. On the Local Traffic Virtual Servers overview screen, adjust the dimensions pane so that it fills up at least half of the screen.
  17. In the Name column, click the magnifying glass icon ( ), type the host name of the BIG-IP device noted in step 8, then click the magnifying glass icon again.
    The chart pane displays statistics for only the selected device. Additionally, only the virtual servers that reside on that device are available on the Virtual Servers dimension.
  18. In the dimensions pane, click the down arrow on the Virtual Servers dimension to expand the list.
  19. In the Bytes Avg/s column, click one of the column headings (for example
    C-->
    ) to sort the list of virtual servers by average traffic level in bytes per second on this virtual server.
  20. Find the virtual server that has the highest number of new connections, and select it.
    The chart pane now displays statistics for the virtual server on the distressed device.
With the ID of the virtual server, you can figure out which application is triggering the traffic spike and figure out what your next step is from there.

Analyze object configuration issues affecting your application

Before you can analyze application object performance data:
  • There must be a BIG-IQ Data Collection Device configured for the BIG-IQ device that manages your BIG-IP devices.
  • It is also a good idea, though not a requirement, to define the retention policy for the statistics data you are collecting.
  • You will need to know the name of the virtual server, pool or pool member that hosts the application that is having performance issues.
When you learn that an application is having performance issues, you can analyze the objects that serve that application to find the likely cause. This example applies to issues that apply to virtual servers. You may also use process to analyze an application's Pools and Nodes (Pools & Pool Members).
  1. At the top of the screen, click
    Configuration
    .
  2. On the left, expand
    LOCAL TRAFFIC
    .
  3. Click
    Virtual Servers
    .
    The Virtual Servers screen opens showing a list of virtual servers managed by this BIG-IQ.
  4. On the Virtual Servers screen, in the Filter in the upper right corner, type the name of the virtual server that is hosting the troubled application.
  5. When you find the virtual server, select the virtual server name.
    The properties screen for the selected virtual server opens.
  6. Click the
    View Statistics
    button, located at the top right of the screen.
    To view statistics for multiple virtual servers, you can select one or more rows in the Virtual Servers screen. Click
    More
    and select
    View Statistics
    The Virtual Servers overview screen opens, but the only items selected, which are the virtual server you selected, its associated applications and the BIG-IP device on which it runs.
  7. Scan the statistics charts plotting data for the distressed server, and look for data throughput trends that indicates a problem (Charts: Throughput Bytes and Throughput Packets).
    There are two ways you can zoom in on an area of interest:
    • Click either end of the focus element on the time selector to define the interval you want to examine.
    • Click and drag over the interval. A small magnifying glass pops up over the area you highlight. When you click the magnifying glass, the time selector focus is set to the area you highlighted.
    If scanning the charts does not reveal an obvious cause, your next step might be to look for a troubled pool member or node.
  8. On the Virtual Servers overview screen, click the back arrow ( ) to return to the properties screen for the virtual server. Click it again to return to the Virtual Servers screen.
  9. Find the virtual server name again, and select the check box that corresponds to it.
    At the bottom of the Virtual Server screen, a two panel preview pane opens displaying information about the virtual server.
  10. On the right panel of the preview pane, under Related Items, click
    Show
    .
  11. Select one of the Pool Members to display its properties screen, and then click
    View Statistics
    .
    The Pools & Pool Members overview screen opens, and again, the chart data that is displayed is only for the BIG-IP, pool, and pool name that you selected.

Identify load balancing issues affecting your application performance

Before you can analyze load balancing data:
  • There must be a BIG-IQ data collection device configured for the BIG-IQ device that manages your BIG-IP devices.
  • It is also a good idea, though not a requirement, to define the retention policy for the statistics data you are collecting.
When you get a report that an application is having performance issues, you can quickly determine if the cause is related to a load balancing problem cause by one of its pool members, pools, or virtual servers. The following example identifies pools members that affect an application's performance.
  1. At the top of the screen, click
    Monitoring
    .
  2. On the left, expand
    DASHBOARDS
    .
  3. Expand
    Local Traffic
    and click
    Pools & Pool Members
    The Pools & Pool Members dashboard opens.
  4. In the dimensions pane, click the down arrow on the
    Applications
    dimension to expand the list.
  5. Find the reported application in the dimension list and select its tow.
    The charts and dimensions shows data that corresponds with the selected application.
  6. Expand the
    Pool Members
    dimension to view the objects that correspond with the selected application.
    You can check the connections and throughput data in the dimensions to evaluate whether there is troubled activity in the transaction process for one or more pool members that host the application.
    You can expand the
    Pool Names
    or
    BIG-IP Host Names
    dimensions to view with objects correspond with the selected application.
  7. Click the name of each pool member until all of the members are selected.
  8. Right click selected names, and select
    Add Comparison Chart
    .
    A new chart is added to the top of the chart pane. The chart plots the performance of the pool members, graphing the load balancing performance over time.
  9. To change the comparison metric that plots for the selected pool members, click the down arrow in the title of the chart and select the new metric.
    You might want to try looking at the average bytes going to or from the server, or maybe the average new server connections.

Troubleshoot DNS traffic issues

Before you can troubleshoot DNS traffic performance issues:
  • There must be a BIG-IQ data collection device configured for the BIG-IQ that manages your BIG-IP devices.
  • It is also a good idea, though not a requirement, to define the retention policy for the statistics data that you are collecting.
When you learn that one of your DNS sync groups is having performance issues, you can analyze the DNS traffic to find the likely source of the issue.
  1. At the top of the screen, click
    Monitoring
    .
  2. On the left, click
    DASHBOARDS
    DNS
    Overview
    .
    The DNS overview screen opens to display dashlets and summary information about your DNS traffic.
  3. Review the dashlets and summary charts for anomalous performance. When you spot a problem, you can find details using the other
    DNS
    options:
    Traffic
    ,
    GSLB
    ,
    Services
    , and
    Attacks and Violations
    . Explore the charts until you spot the most likely source of the performance issue.