Interface metrics show gaps in DX NetOps Performance Management reports

book

Article ID: 106224

calendar_today

Updated On:

Products

CA Infrastructure Management CA Infrastructure Management CA Performance Management - Usage and Administration DX NetOps

Issue/Introduction

Generating Report Interfaces appears Gaps independently of Time Rage reports

How to fix the Gaps in Interfaces reports

When running reports for Interfaces often gaps in the data are observed. The gaps are intermittent, often random, but occur often.

The metrics the gaps are seen for vary, often for Bits In/Out metrics or other metrics based on the Speed configured on the Interface at issue.

Cause

There are various possible causes for the gaps in the Interface items metric data. The common causes are:

  • The Interfaces Device or it's Component Item did not respond to a poll request. We could have timed out waiting for the response, or it might have sent back a different error.
  • The Device is down
  • The Interface Component Item is:
    • Set to ifAdminStatus of DOWN
    • No longer present at it's previously known ifIndex Index value.
  • A network problem causes the poll response to be lost, not returned in a timely manner (timeout) or not returned at all.
  • The Data Collector/Aggregator is down or having problems.
  • The MIB OID counters rolled over because of:
    • Incorrect Vendor Certification Priorities order resulted in use of incorrect Vendor Certification
    • The Interfaces ifSpeed and/or ifHighSpeed is not set or set incorrectly.

Environment

All supported DX NetOps Performance Management releases

Resolution

The resolution for these various causes are equally different.

  • For most instances it will require review of the logging on the Data Collector managing the affected item and it's device to determine the cause.
    • Gather a diagnostics logging package via the re.sh script from the Data Collector involved.
    • Attach that logging package to a new Support Case seeking input.
    • Reference this Knowledge Base Article in the new Support Case.
    • Instructions for the re.sh script are found in the Unable to Resolve Issue documentation.
  • If the Device is down, have it restarted and once resolved and restarted, see if the issue in reports is resolved.
  • If the Interface is set to ifAdminStatus down, have it set to Up. By default the Interface Metric Family has a filter that prevents polling of Interfaces set to ifAdminStatus=DOWN.
  • If the Interface has changed the index it is assigned in the MIB data, run a Metric Family Update against the Interface Metric Family for affected devices.
    • Done via the Polled Metric Families tab for the Device in the Data Aggregator Monitored Devices Inventory.
  • If it's a network problem the best way to identify the cause is via a packet capture using tcpdump, snoop or similar tools.
  • For Counter Rollover problems:
    • If the wrong Vendor Certification is in use change the Vendor Certification Priorities order for the Interface Metric Family.
      • Set the wrong one to a higher number in the list than the correct one.
      • This will trigger a rediscovery. When it completes it should use the correct Vendor Certification instead of the wrong one, if the device supports the MIB OIDs in the correct Vendor Certification.
    • If the Interface is using the default High Speed Interface Vendor Certification, many metrics rely on the correct ifSpeed and/or ifHighSpeed settings for the interface to correctly calculate the metrics. If the Speed is set incorrectly or is not set at all, we'll see missing metric data or metric data gaps. To resolve this:

Additional Information

Can also review the documentation regarding Gaps Appear in Reports or Views.