Enabling the event counter extension

Functionality delivered in fix pack
18The event counter extension provides the tools and configuration that is required to gather the metrics for the IBM Netcool Operations Insight licensing model, which is based on event count.

Before you begin

The event counter relies on some of the probe self-monitoring configurations that can be found in the $NCHOME/omnibus/extensions/roi folder. For more information, see Configuring probes for self monitoring.
  • ObjectServer table and triggers are defined in the $NCHOME/omnibus/extensions/roi/probestats.sql file.
  • Probe include rules file: $NCHOME/omnibus/extensions/roi/probewatch.include

About this task

Individual probes need to be configured so that they gather their own statistics at fixed intervals and then forward them on to the ObjectServer. Triggers in the ObjectServer aggregate the event count from the combined probe feeds.

The default configuration sets the statistics gathering according to the following parameters:
  • The probes forward their statistics to the ObjectServer once every minute
  • The ObjectServer calculates the total events every 10 minutes.
  • The ObjectServer stores this data for up to 2 years.

Housekeeping triggers are provided to prune the data periodically, which prevents the table from growing indefinitely.

A stored procedure is provided, which when called, creates a report that lists the event totals by calendar month. For more information, see Creating a report for event totals by calendar month.

Events received or processed by a probe between the last heartbeat and the time when the probe goes down, which is determined by the ProbeWatchHeartbeatInterval parameter, are not added to the overall event count. This issue occurs because the probe does not send a final stats event to the ObjectServer before it shuts down.

Event counts are collected per ObjectServer. In a tiered architecture they are not propagated to the aggregation server.

Event count statistics are not kept in sync between a failover pair.

To enable the event counter extension, complete the following steps:

Procedure

  1. Note: To load the Event Counter extension, a user must be a member of the System group in the ObjectServer.
    Load probe statistics support into the ObjectServer (if not already installed). Run one of the following commands, depending on your operating system:
    For UNIX operating system
    $NCHOME/omnibus/bin/nco_sql -server servername -user username -password password < $NCHOME/omnibus/extensions/roi/probestats.sql
    For Windows operating system
    %NCHOME%\omnibus\bin\isql.bat -S servername -U username -P password -i %NCHOME%\omnibus\etc\extensions\roi\probestats.sql
  2. Load event counter triggers and the procedure:
    For UNIX operating system
    $NCHOME/omnibus/bin/nco_sql -server servername -user username -password password < $NCHOME/omnibus/extensions/eventcounter/eventcounter.sql
    For Windows operating system
    %NCHOME%\omnibus\bin\isql.bat -S servername -U username -P password i %NCHOME%\omnibus\extensions\eventcounter\eventcounter.sql
  3. Activate housekeeping on the probe statistics table by enabling the probe_statistics_cleanup trigger:

    GUI Use the Administrator GUI panel to simply enable the trigger. Go to Automation > Triggers. For more information, see Creating and editing database triggers.

    For UNIX operating system
    $NCHOME/omnibus/bin/nco_sql -server servername -user username -password password
    > alter trigger probe_statistics_cleanup set enabled TRUE;
    > go
    For Windows operating system
    %NCHOME%\omnibus\bin\isql.bat -S servername -U username -P password 
    > alter trigger probe_statistics_cleanup set enabled TRUE;
    > go
  4. Configure the probe rules file:
    1. Copy the $NCHOME/omnibus/extensions/roi/probewatch.include file to the $NCHOME/omnibus/etc/rules/ folder, or to your preferred local or remote directory, where the main rules file or any secondary rules files for the probe is stored.
    2. Edit the $NCHOME/omnibus/etc/rules/probewatch.include file. Add registertarget(), as described in the MODIFICATION REQUIRED comment in the file. Add the corresponding genevent() to send statistics events to the defined target server.
    3. Replace the entire ProbeWatch section with code similar to the following Simnet Probe example:
      include “$NCHOME/omnibus/etc/rules/probewatch.include”
      
      @Manager = "Simnet Probe"
      @Class = 3300
      @Node = $Node
      @Agent = $Agent
      @AlertGroup = $Group
      @Summary = $Summary
      @Severity = $Severity
      @Identifier = $Node + $Agent + $Severity + $Group
      if (nmatch($Summary, "Port failure"))
      {
           @AlertKey = $PortNumber
      }
      else if (nmatch($Summary, "Diskspace"))
      {
           @AlertKey = $PercentFull + "% full"
      }
      if (regmatch($Summary, ".*Down.*"))
      {
           @Identifier = "Down" + @Identifier
           @Type = 1
      }
      if (regmatch($Summary, ".*Up.*"))
      {
           @Type = 2
      }
    4. Optional: If required, edit the new ProbeWatch section in the include file and reinstate any preexisting customizations.
    5. Edit the $NCHOME/omnibus/etc/rules/probewatch.include file to enable the following parameters:
      $OplHeartbeat_discard=1
      $OplHeartbeat_populate_master_probestats=1
    6. The following probe property should also be set in the probe’s propsfile:
      ProbeWatchHeartbeatInterval: 60 # Collect probe stats once every minute