This documentation supports the 22.1 version of BMC Helix ITSM Insights.

To view an earlier version, select the version from the Product version menu.

Configuring the recurrent job settings for proactive problem management

As a problem coordinator, you can configure recurrent clustering jobs for proactive problem management based on the specific requirements of your organization. You can create a maximum of three recurrent jobs.
When you set up recurrent jobs, the algorithm runs at the selected interval to cluster incidents based on the selected criteria.

By default, a maximum of 200000 incidents are considered while creating a recurring Proactive problem management job. However, you can increase the limit by customization.

To create a recurrent job

  1. Click Manage Job.
  2. On the Proactive problem management settings page, click Create recurrent job
  3. In the General section, enter a job name. 
  4. Select the language for the incident text processing. Even though your incident text contains mixed language, pre-processing is done on the basis of the selected language.

  5. In the Data Set section, you can specify the data fields on which you can create clusters and filter data. Select the fields on which you want to create clusters. 

    To also view the required system fields, select the Show fields required by the system check box. The system fields cannot be removed. Except for Submitter, all other system fields cannot be removed. Some fields may be hidden from the data set by your admin to comply with privacy regulations.

  6. You can specify filters on the fields to further refine your data set. 

    1. Click the required filter category and select one of the following options:

      OptionDescription
      Equals to

      Select this option to include values in the filter.

      Not equals to

      Select this option to exclude values from the filter.

    2. Search and select the field value that you want to include or exclude.
    3. Click Apply filters.

      • Searching for a string without a wildcard (%) is not supported in a filter that has a text field. We recommend using a wildcard (%) for a search in such filters.
      • The Equals to and Not equals to options appear only in fields with a character menu, such as Service CI.


  7. In the Schedule section, you can toggle to enable or disable the scheduled job run. You can configure a maximum of three recurrent jobs. 
    • Specify how often the job needs to run: daily, monthly, or weekly. The default value is monthly.
    • For a weekly job, specify the day of the week on which the job should run. By default, this value is Monday.

    • For a monthly job, specify the day of the month on which the job should run and select every n month to indicate frequency of the job run. For example, if you select 2, it means that the job is to be run every two months. 

    • Specify the date from which the job is to be run. You can also specify an end date or leave it blank.
    • Specify the look back period. Look back time period is the period used to search for incident data for a job run. For example, a look back period of one month means that the job uses a month's data until the day the job is run.

      Best practice

      Define your schedule based on the problem management process in your organization. The schedule for clustering jobs can be weekly or bi-weekly or monthly based on your review cycles for problem identification. Typically, the lookback period for each of these schedules is the previous 1-4 weeks of data.

    • Select Look back date field, which is the date field that is used to search for incidents within the look back time period.
  8. In the Create clusters section, specify the parameters by which the data is to be grouped. 

    1. For the first level of grouping, select from one of the data fields. If you select Machine learning, you cannot select the second level grouping. 
    2. For the second level of grouping, you can select Machine learning or another set of data fields (excluding the data field already selected in level 1 grouping). 
    3. If you have selected Machine learning, you can specify the text fields that will be clustered. Summary is the default text field used for clustering. If more than one text field is provided, these fields are concatenated into one field.

      Important

      If you have not selected Machine Learning in any of the Group by fields, then a clustering job is not run and incidents are grouped by the selected fields.

  9. In the Advanced machine learning section, you can specify the number of clusters. To allow the system to set the number of clusters, click the Let the system set the number of clusters check box.

    Best practice

    When selecting the number of clusters, we recommend selecting the Let the system find no. of clusters check box rather than setting the number of clusters yourself. The system automatically selects an optimal number of clusters. However, when you know the number of clusters from prior execution runs or domain knowledge, you can specify a value to improve the response time. We recommend setting 20-30 clusters for optimal incident monitoring.


  10. You can upload custom stop words in a .txt file, where each stop word is defined on a new line. Every time you upload a new file, it overrides the old file and the stop words in the new file are used.
    BMC Helix ITSM Insights has a built-in library of stop words for every supported language. The algorithm refers to the library and your preferred stop words while processing incident information.

    Important

    While generating cluster labels in the dashboard from the relevant incidents, the algorithm compares the incident description words with the stop word library. Therefore, the cluster labels do not contain words mentioned in the stop word library.

    • from
    • subject
    • re
    • edu
    • use
    • need
    • able
    • bmc
    • com
    • abc
    • however
    • ourselves
    • alone
    • us
    • would
    • already
    • most
    • off
    • back
    • and
    • none
    • because
    • where
    • first
    • it
    • nevertheless
    • too
    • each
    • whereupon
    • wherein
    • as
    • this
    • whatever
    • always
    • serious
    • then
    • ‘ve
    • not
    • he
    • them
    • n't
    • even
    • thru
    • anyway
    • above
    • eight
    • 'm
    • or
    • besides
    • hereby
    • than
    • during
    • being
    • never
    • therein
    • has
    • does
    • hereupon
    • whereafter
    • is
    • becomes
    • ca
    • get
    • seemed
    • nowhere
    • nobody
    • rather
    • whenever
    • yourselves
    • few
    • may
    • elsewhere
    • but
    • my
    • again
    • will
    • 're
    • more
    • once
    • 'm
    • otherwise
    • anything
    • various
    • had
    • together
    • within
    • via
    • are
    • fifty
    • afterwards
    • mine
    • how
    • many
    • thereupon
    • should
    • himself
    • everything
    • against
    • sixty
    • perhaps
    • although
    • 's
    • along
    • except
    • his
    • whether
    • anywhere
    • must
    • one
    • their
    • s'
    • for
    • no
    • someone
    • upon
    • meanwhile
    • might
    • here
    • namely
    • indeed
    • under
    • n‘t
    • almost
    • least
    • forty
    • we
    • everyone
    • toward
    • before
    • if
    • show
    • about
    • please
    • another
    • through
    • 've
    • to
    • unless
    • side
    • also
    • move
    • any
    • can
    • just
    • two
    • throughout
    • three
    • could
    • me
    • across
    • whence
    • else
    • five
    • amongst
    • mostly
    • wherever
    • four
    • hence
    • front
    • anyone
    • herself
    • whereby
    • somehow
    • whose
    • ever
    • go
    • was
    • themselves
    • since
    • when
    • noone
    • 's
    • with
    • same
    • did
    • n’t
    • per
    • yourself
    • am
    • become
    • beside
    • well
    • why
    • they
    • 'll
    • regarding
    • ours
    • her
    • give
    • our
    • made
    • ‘re
    • thereby
    • an
    • much
    • full
    • ten
    • take
    • 're
    • out
    • such
    • therefore
    • over
    • still
    • seem
    • 'd
    • former
    • latter
    • next
    • everywhere
    • ‘d
    • quite
    • becoming
    • amount
    • sometime
    • twenty
    • doing
    • last
    • name
    • ’ll
    • whom
    • these
    • every
    • latterly
    • make
    • seeming
    • among
    • part
    • that
    • twelve
    • either
    • of
    • myself
    • say
    • hundred
    • thereafter
    • those
    • at
    • you
    • down
    • several
    • herein
    • been
    • beyond
    • nine
    • him
    • towards
    • what
    • onto
    • do
    • both
    • sometimes
    • thence
    • moreover
    • the
    • due
    • beforehand
    • empty
    • fifteen
    • thus
    • anyhow
    • have
    • whereas
    • eleven
    • done
    • yours
    • she
    • your
    • whoever
    • who
    • only
    • own
    • somewhere
    • formerly
    • between
    • others
    • neither
    • below
    • up
    • its
    • hers
    • be
    • whither
    • were
    • into
    • yet
    • became
    • ‘m
    • less
    • until
    • all
    • see
    • used
    • seems
    • now
    • enough
    • often
    • call
    • without
    • on
    • a
    • something
    • further
    • put
    • though
    • after
    • while
    • bottom
    • six
    • nothing
    • hereafter
    • behind
    • which
    • very
    • so
    • top
    • using
    • whole
    • i
    • there
    • really
    • in
    • ’d
    • nor
    • third
    • cannot
    • around
    • itself
    • by
    • keep
    • other
    • ’ve
    • ‘ll
    • some

    The following table describes the usage of in stop words:

    Incident summaryStop wordDescription
    ITSMInsights is running low on memoryITSM%Removes the stop word ITSM and the characters following it.
    In this case, ITSMInsights is removed from the resulting cluster label.
    ITSMInsights is running low on memory%Ins%Removes the stop word Ins and the characters preceding and following it.
    In this case, ITSMInsights is removed from the resulting cluster label.
    ITSMInsights is running low on memory%InsightsRemoves the stop word Insights and the characters preceding it.
    In this case, ITSMInsights is removed from the resulting cluster label.

    Important

    • Stop words are case-sensitive. We recommend matching the case of your stop word with that of the word in the description that you want to remove.
    • If you configure custom priority values after creating a recurring job, the relevant clusters may not appear in the Proactive problem management dashboard. Therefore, we recommend you edit and rerun the existing recurrent jobs after you configure custom priority values.
  11. Click Save.

The job configuration is saved. A job might take several minutes to complete, depending on the incident data to be processed. Refresh the jobs table to check if the job is completed. If multiple problem coordinators schedule jobs to be run at the same time, the jobs might get queued. 

When the job run is completed, the jobs table displays the job status in the Jobs table. If the job run was not successful, the Last job message column displays the reason why the job failed. Once the job is successfully run, you can select it from the Jobs list in the dashboard to view the clusters.

Important

  • Depending on number of records or incidents and the kind of incident data, the number of clusters on the dashboard could be less than the number you have specified.
  • Running a job multiple times with the same parameters may generate a different number of clusters. You may notice a slight difference in the number of clusters generated under the following conditions:
    • When you use Group by fields and let the system generate the ideal number of clusters.
    • When you use Group by fields and provide a cluster value.
  • The algorithm groups incidents that have similar words in their description and avoids generating one-word cluster labels in the dashboard. Therefore, if an incident has a one-word description, the algorithm finds an existing cluster that has incidents with that word in its description sentence and adds the incident to that cluster.

To edit a recurrent job

To edit a recurrent job, click the edit icon in the Actions column. Make the necessary changes to the job.

The changes you have done will take effect in the next scheduled job.

To delete a recurrent job

To delete a recurrent job, click the delete icon in the Actions column. When you delete a job, the job definitions, that is, the data fields and filters applied are also deleted. Also, all job runs associated with that job are deleted.

Was this page helpful? Yes No Submitting... Thank you

Comments