Poll Critical Interfaces Faster than Non-critical Interfaces

As the Administrator, you need frequent data about your most critical systems while maximizing the overall performance of your performance management systems. One way to accomplish your goal is by polling only critical interfaces at a high rate, while polling noncritical interfaces at a normal or slow rate. You can poll at differing rates by using a filter on the Interfaces metric family that is associated with your monitoring profile. By fast-polling interfaces sparingly, you can reduce unnecessary network traffic and performance management system load while still sufficiently monitoring the health of your network system.
capm370
As the Administrator, you need frequent data about your most critical systems while maximizing the overall performance of your performance management systems. One way to accomplish your goal is by polling only critical interfaces at a high rate, while polling noncritical interfaces at a normal or slow rate. You can poll at differing rates by using a filter on the Interfaces metric family that is associated with your monitoring profile. By fast-polling interfaces sparingly, you can reduce unnecessary network traffic and performance management system load while still sufficiently monitoring the health of your network system.
For example, your data center access switch connects many application servers to only two aggregation switches. You decide to poll the interfaces supporting these aggregation switches at a higher rate. These links are critical, because they support network traffic to all other connected switches. However, polling
all
interfaces at a higher rate would cause unnecessary network traffic, wasting system resources and possibly causing network performance issues. After consulting with your network operations and engineering teams, you decide that a normal polling rate is sufficient for the interfaces connecting each attached server. To apply different polling rates, you implement two monitoring profiles for interfaces.
Filters that you set on metric families are ignored when event rules that are applied to monitoring profiles trigger events.
2
View Your Monitoring Profiles
As the CA
CA Performance Center
Administrator, you decide to poll critical interfaces as often as possible. However, you want to minimize unnecessary network traffic that polling
all
interfaces at this fast rate can produce. You decide to create two monitoring profiles for interfaces -- one with normal polling, and one with fast polling.
Before you create a monitoring profile, you review the existing monitoring profiles to find one that closely matches your needs.
Follow these steps:
  1. Click Monitoring Profiles from the Monitoring Configuration menu for your Data Aggregator data source.
    A list of monitoring profiles is populated.
  2. Select a monitoring profile.
    Details for the selected monitoring profile populate the tabs:
    • Metric Families tab -- Displays a list of metric families that are associated with that specific monitoring profile. Metric families contain the metrics that are used for polling devices and components.
    • Collections tab -- Displays a list of device collections that are associated with that specific monitoring profile.
Copy a Factory Monitoring Profile
As the CA
CA Performance Center
Administrator, you find that the factory Network Interfaces monitoring profile closely matches your needs and requires only minor changes. Therefore, you create a copy and use it to poll only critical interfaces at a faster polling rate.
Log in as the administrator to perform this task.
Follow these steps:
  1. Navigate to the list of all monitoring profiles in CA
    CA Performance Center
    .
  2. Select the Network Interfaces monitoring profile and click Copy.
    Factory monitoring profiles cannot be edited or deleted. All monitoring profiles, including custom, are global.
    The Create/Edit Monitoring Profile dialog opens.
  3. Enter the following information for your monitoring profile:
    • Name:
      Uplink Interfaces
    • Description
      (optional)
      :
      Monitors performance of interfaces in all critical Uplink devices.
    • SNMP Poll Rate:
      1 minute
    We recommend that you rename the profile. Unique naming is enforced across all tenants.
    Consider the following information about poll rates:
    • When the poll rate is changed, it takes up to two cycles for the new poll rate to take effect. When the 60-minute rate is used to poll an existing device, a 'No Data To Display' message appears in the dashboard view given the default time range of Last Hour. If you change the dashboard setting to a prior hour, it is possible to see earlier data. However, the view does not display the latest data until the new poll cycle completes.
    • Interfaces that are assigned to multiple monitoring profiles with different poll rates are polled at the fastest assigned rate.
  4. Leave the Change Detection Settings, Rate value at 24 Hours.
    Consider the following information about change detection rates:
    • The
      change detection rate
      is how often Data Aggregator checks whether any components on a device have been reconfigured. Changes can include new components that have been created or existing components that have been retired.
      The reconciliation algorithm specified in the metric family defines the configuration changes to watch for.
    • The Change Detection Settings, Rate option is used to set the frequency at which Data Aggregator checks for changes. The rate of detection can be set in minutes or hours. By default, the rate is set to 24 hours.
    • Changes are detected at the fastest rate you specified for all of the monitoring profiles that are associated with a collection of devices.
  5. Leave the 'Automatically Update Metric Families' check box selected.
    This option controls the Data Aggregator response once a change or reconfiguration is detected. Selecting this option automatically causes Data Aggregator to start monitoring new components or to stop monitoring retired components. When this option is not selected, you can manually control monitoring of components, as follows:
    1. Manually check the Events Display dashboard to watch for configuration events.
    2. Navigate to the Data Aggregator administration menu, Monitored Devices, Polled Metric Families view.
    3. Select the appropriate metric family, and click Update Metric Family to help ensure that Data Aggregator picks up the latest device reconfiguration.
    If an interface filter is applied, Data Aggregator monitors only the interfaces that pass the filter conditions after reconfiguration.
  6. Leave the Interfaces metric family as the only metric family in the Selected Metric Families list.
  7. Click Save.
    Your copied monitoring profile is added to the Monitoring Profiles list. However, this monitoring profile is not active until you assign it to a device collection.
Set an Interface Filter
By default, the factory network interface monitoring profile includes a filter to prevent modeling interfaces that are administratively down. In addition, interfaces with a type (ifType) of 1 (Other) or 24 (Loopback) are not modeled, regardless if those interfaces are administratively up or down. IPSLAs with the rttMonCtrlAdminOwner MIB object that contains the string “Network Health" are not modeled either.
Filtering reduces the number of interfaces that are monitored, which reduces unwanted data collection and network traffic.
In addition to polling only administratively up interfaces, you also want to poll the most critical interfaces more frequently. To isolate and poll only these interfaces faster, you add a second filter condition to the interface filter associated with your custom monitoring profile. This second filter condition isolates the critical interfaces by finding only interfaces that contain "uplink" in their description.
Log in as the administrator to perform this task.
Follow these steps:
  1. Select your interfaces monitoring profile (called "Uplink Interfaces") from the Monitoring Profiles page.
  2. Click Interface metric family row on the Metric Families tab and click Edit Filter.
    Do not click directly on the metric family name, because it is linked to take you to the metric family definition. Instead, click the row the metric family name is in to activate the Edit Filter option.
  3. Click the Add Condition button.
    Multiple conditions are connected with an "and" operation. That is, all conditions must be met to satisfy the filter.
  4. Configure the filter conditions with the following options and click Save:
    • Attribute: Description
    • Operation: Contains
    • Filter Value: uplink
    The Filter Value field is case-sensitive.
    Consider the following details about additional attributes you can use for filtering:
    • For Speed In and Speed Out, you can use a decimal in the text field (such as 1.544) and can specify bps, Kbps, Mbps, or Gbps.
    • For more information about configuring Type (that is, ifType), see the iana web site: .
    • For Description and Alias, you can use a regular expression for filtering only when you select the Matches Regex or the Does Not Match Regex operation.
    • When you save your changes, the filter criteria display on the Metric Families tab. You can now apply this monitoring profile to the appropriate device collection to begin polling your selected interfaces.
     
Data Aggregator applies filtering after discovery. Interface components that do not match the filter criteria are not polled. If you add or edit an Interface filter
after
you run a discovery, polling on these components stops. These interface components are
not
displayed in CA
CA Performance Center
dashboards and data views.
Considerations for Interface Filters and Multiple Monitoring Profiles
When multiple monitoring profiles are assigned to a device collection, the filter matching criteria follows the "or" rule. So, Data Aggregator monitors all interfaces that satisfy the criteria for any of the monitoring profiles in the group.
Some of the monitoring profiles may have filters and some may not. Plus, these profiles can specify differing poll rates. In this case, Data Aggregator monitors the interfaces that match any monitoring profile, but the polling rates can differ. If more than one monitoring profile applies to an interface, Data Aggregator polls the interface once, and polls it at the fastest polling rate:
  • Monitoring Profile 1 -- Filter: Description contains "X," Poll Rate: 1 minute
  • Monitoring Profile 2 -- Filter: None, Poll Rate: 5 minutes
  • Monitoring Profile 3 -- Filter: Description contains "Y," Poll Rate: 10 minutes
In this example, interfaces that match Monitoring Profile 1 are polled every minute. All other interfaces are polled every 5 minutes. Interfaces that match Monitoring Profile 3 also match Monitoring Profile 2, which does not include a filter. The fastest poll rate applies, so no interfaces are polled at 10-minute intervals.
In this case, if one monitoring profile has no filter, the result is that many interfaces may be polled more frequently than necessary. Therefore, after you set a filter, remove associations from other monitoring profiles to make sure that only components matching the specified filter are monitored.
Assign Your Monitoring Profile to a Device Collection
As the administrator or a tenant administrator, you associate the new Uplink Interfaces monitoring profile with a device collection to begin polling. In this case, you associate the profile with the Switches device collection, which is the same device collection that is associated with the factory Network Interfaces monitoring profile. Polling rates are applied to the interfaces in this device collection, as follows:
  • Fast polling: Interfaces that satisfy the filter criteria of the Uplink Interfaces monitoring profile.
  • Normal polling: All other interfaces that the Network Interfaces monitoring profile discovers.
All custom monitoring profiles are global and visible to tenant administrators. However, the association of a monitoring profile with a specific device collection can be scoped to a tenant.
Follow these steps:
  1. Click Collections from the Monitoring Configuration menu for your Data Aggregator data source.
    A list of device collections displays. Administrators can view the device collections for the tenant they are administering. A tenant administrator can view its own (tenant) list of device collections.
  2. Select the All Switches device collection and click the Monitoring Profiles tab.
    A list displays the monitoring profiles that are associated with the selected device collection. The Network Interface device collection exists in this list.
  3. Click Manage.
    The Assign Collection Monitoring Profiles dialog opens.
  4. Select the Uplink Interfaces monitoring profile and click Add.
    The selected monitoring profile moves to the Assigned Monitoring Profiles list.
  5. Click Save.
    Your changes are saved.
View Monitored Devices to Verify Results
After you set up your monitoring profiles, review the monitored devices and the Filter report to verify that only your critical devices are polled at the higher rate. This information helps you to see information in context, such as which monitoring profiles are being used to poll device components. Verifying the results can help you identify any necessary adjustments to help you achieve the polling results that you want.
Note:
Monitored devices are manageable devices and pingable (accessible but not manageable). Inaccessible devices are not monitored devices. Components of monitored devices can be viewed from the Polled Metric Families tab.
Follow these steps:
  1. Run an on-demand discovery.
    If your discovery profile runs automatically, you can wait for the next scheduled discovery.
  2. Click Monitored Devices from the Monitored Inventory menu for a Data Aggregator data source.
  3. Select one of these options from the drop-down list to locate one of your aggregation switch devices in the corresponding tree view:
    • Device by Collection -- Your devices appear under the All Switches device collection.
    • Device by Monitoring Profile -- Your critical interfaces appear under Devices under the Uplink Interfaces monitoring profile.
    Alternatively, select the Search tab to search by host name, device name, or IP address. You can enter a partial name or IP address to return a list of devices that contain that partial match. Wildcards and regular expressions are not supported.
    The Polled Metric Families tab shows the consolidated monitoring profiles that are associated with the switch device. Devices have only one consolidated monitoring profile. Each consolidated monitoring profile lists every metric family to poll on the device and whether the device supports the metric family.
  4. Select the Interface metric family.
    The Components table for the Interfaces metric family shows one of the following polling statuses for the discovered Interface components:
    • Active
      Indicates that the component is being polled.
    • Inactive
      Indicates that polling has stopped on the component because the metric family is no longer monitored for the device.
    • Not Present
      Indicates that the component no longer exists on the physical device. Polling is stopped on the component. You can view historical data for reporting purposes. By default, retired components are not synchronized with CA
      CA Performance Center
      .
    • Filtered (interface components only)
      Indicates that the component does not pass the filter criteria and polling on the component is stopped.
      Filtered interfaces are not displayed in CA
      CA Performance Center
      dashboards and data views.
       
  5. (Optional) Select the Interface metric family and click Update Metric Family.
    Data Aggregator reconfigures components for any configuration updates. For example, if you add a disk drive on a server, you can use the Update Metric Family button to rediscover the configuration update. The configuration update creates a disk component.
  6. Click the Filter Report tab and follow these steps:
  7. Look at the filters on each of the other Interface monitoring profiles to see if they are monitoring the same device collection that you want to filter.
  8. Remove any relationships between other Interface monitoring profiles and device collections that will block your filter criteria. For example, if your new Interface monitoring profile is associated with the All Routers device collection, remove the relationship between
    other
    Interface monitoring profiles and the All Routers device collection.
  9. Run another discovery and review the updated Filter report to verify that the new filter criteria is active. If the Filter report shows that an unwanted monitoring profile was included, repeat the previous steps until you are monitoring only the interfaces that you want.
    The Filter Report tab shows which interface filter criteria have been used during component monitoring. The tab also shows a report of all of the interfaces that are identified on the device and whether they matched the specified filter criteria.
If you change the rules on a custom monitoring profile, the Interface Filter Criteria pane does not reflect those changes. If you disassociate the monitoring profile from a group, the Interface Filter Criteria pane does not reflect those changes. Rediscover the device to filter the interfaces that are based on the changes you made to the filter criteria and monitoring profile.