Settings

Prev Next

The Settings page allows administrators to configure how Landscape Intelligence (LI) processes incident data, identifies incident clusters, detects recurring problems and incident spikes, generates resolution suggestions, and identifies potential service request candidates.


Figure: Settings

The Configuration page is organized into the following tabs:

General

The General tab defines the cluster configuration and historical data window used by Landscape Intelligence scans.


Figure: General

Setting

Description

Minimum cluster size

Specifies the minimum number of incidents required for the Diagnostic Run to accept a group as a cluster. Increase this value to reduce noisy data and generate fewer, more consolidated clusters.

Minimum candidate size

Specifies the minimum number of incidents required for a theme to be considered a Problem Record candidate.

Lookback window

Specifies how much recent incident history each scan analyzes. This setting is shared by the problem detection and storm alert scans.

Schedules

The Schedules tab allows administrators to configure when the Landscape Intelligence pipeline runs. It includes separate schedules for Diagnostic Run, Problem detection, and Storm alerts.

The schedule times are based on the user's local time zone. The scheduler checks every five minutes and starts the job at the first available tick at or after the configured time.

Figure: Schedules

Diagnostic Run

The Diagnostic Run rebuilds incident clusters and centroids, regenerates cluster reports, and recomputes resolution patterns.

Configure the following options:

Setting

Description

Enable/Disable

Turns the Diagnostic Run schedule on or off.

Frequency

Specifies how often the Diagnostic Run executes, such as daily or weekly.

Weekday (local)

Specifies the day of the week on which the run executes when the weekly frequency is selected.

Run hour (local)

Specifies the hour at which the run executes, based on the local time zone.

Run minute (local)

Specifies the minute at which the run executes.

Problem detection

The Problem detection schedule scans the configured rolling history window for recurring or outlier incident themes and proposes Problem Records.

Configure the following options:

Setting

Description

Enable/Disable

Turns the Problem detection schedule on or off.

Frequency

Specifies how often the Problem detection scan executes.

Run hour (local)

Specifies the hour at which the scan executes.

Run minute (local)

Specifies the minute at which the scan executes.

Storm alerts

The Storm alerts schedule identifies abnormal incident spikes against each cluster's historical baseline.

Configure the following options:

Setting

Description

Enable/Disable

Turns the Storm alerts schedule on or off.

Frequency

Specifies how often the Storm alerts scan executes.

Run hour (local)

Specifies the hour at which the scan executes.

Run minute (local)

Specifies the minute at which the scan executes.

Resolution suggestion schedule

Resolution patterns are recomputed at the end of every Diagnostic Run. There is no separate schedule or enable/disable switch for resolution suggestions.

Problem detection

The Problem detection tab allows administrators to configure the thresholds used by the problem detection scan to identify recurring incident themes that may qualify for Problem Records.

Figure: Problem detection

Configure the following settings:

Setting

Description

Assignment threshold

Specifies the minimum similarity an incident must have with an existing cluster to be assigned to that cluster during problem detection.

Min avg pairwise similarity

Specifies the minimum average similarity required among the incidents in a candidate group before the group is proposed as a Problem Record candidate.

Storm alerts

The Storm alerts tab allows administrators to configure the thresholds used to identify abnormal increases in incident volume within a cluster.


Figure: Store alerts

Configure the following settings:

Setting

Description

Z-score threshold

Specifies how many standard deviations the incident volume must exceed the cluster's baseline before an alert is raised. A lower value makes the alert more sensitive.

Min history days

Specifies the minimum number of days of historical data required before a cluster's baseline is considered reliable enough to generate alerts.

Resolution suggestion

The Resolution suggestion tab allows administrators to configure the confidence thresholds and data requirements for displaying historical resolution patterns as suggestions.

Resolution patterns are recomputed at the end of every Diagnostic Run. They are available only while the resolution notes continue to pass the data-quality audit.

Figure: Resolution suggestion

Configure the following settings:

Setting

Description

Membership strength

Specifies how closely an incident must match a cluster before any resolution suggestion is displayed.

Pattern similarity

Specifies how closely an incident must match a specific resolution pattern to receive that pattern as a suggestion instead of the cluster-wide fallback.

Minimum supporting incidents

Specifies the minimum number of historical incidents required to support a resolution pattern. Patterns with fewer supporting incidents are grouped into the scattered category.

Review traffic threshold

Specifies the incident traffic threshold above which a resolution pattern must be reviewed by a human before it is published.

Shortest usable note

Specifies the minimum length of a resolution note to be considered usable. Notes shorter than this value are treated as sign-off text rather than meaningful resolution information. Leave this field empty to let the data-quality audit determine the threshold automatically.


Early trigger

The Early trigger tab allows administrators to configure an adaptive trigger that starts a Diagnostic Run ahead of its regular schedule when new incident patterns appear faster than the scheduled run can detect them.


Figure: Early trigger

Configure the following settings:

Setting

Description

Adaptive early trigger

Enables or disables the adaptive early trigger. When enabled, it is checked once a day as part of Incident Watch. It does not create a separate job or change the Diagnostic Run schedule.

Novel incident rate

Specifies the percentage of daily incident volume that must be classified as novel or unassigned to any cluster before the early trigger condition is met. A lower value causes the trigger to activate sooner.

Consecutive days

Specifies the number of consecutive days for which the novel incident rate must remain above the configured threshold before an early run is triggered.

Cooldown

Specifies the minimum waiting period after an early Diagnostic Run before another early run can be triggered.

The configuration summary displays the conditions under which an early Diagnostic Run will be triggered.

For example, with the default configuration, an early run is triggered when more than 15% of a day's incident volume is novel or unassigned for three consecutive days. After the run, the system waits 48 hours before another early run can be triggered.

Service requests

The Service requests tab allows administrators to configure the criteria for identifying recurring incident groups that may be suitable for conversion into service catalog items. It also defines when a matching new incident can be converted into a service request.

Candidates are recalculated during every Diagnostic Run using incident clusters and resolution patterns. Changes apply from the next run. A service catalog item is not created automatically; it requires approval from a catalog owner.

Figure: Service Request

Configure the following settings:

Setting

Description

Minimum volume

Specifies the minimum average number of tickets per month required for an incident group to be proposed as a service catalog item. Groups below this threshold are listed but are not proposed.

Months with tickets

Specifies how many of the last three months must contain tickets for a candidate to qualify. This helps prevent one-time incident spikes, such as migration-related bursts, from being proposed as recurring service requests.

Minimum evidence

Specifies the minimum number of historical tickets with a recognized fix, identified from resolution patterns or resolution notes, required to evaluate a request candidate.

Request evidence

Specifies the minimum percentage of recognized fixes that must represent fulfillment, information, or a change for the candidate to qualify as a service request. Fixes classified as repairs remain incidents.

Automation tier

Specifies the percentage of one type of fix that must be represented before the candidate receives an automation tier, such as on its own, prepare and approve, or agent-assisted.

Convert at logging

Specifies the minimum percentage of fulfillment fixes required before a matching new ticket is automatically converted into a service request. Below this threshold, the dispatcher receives a suggestion instead.

The configuration summary provides an overview of the criteria used to propose catalog items and convert matching incidents.

For example, with the default configuration, a candidate is proposed when:

  • It averages at least five tickets per month.

  • Tickets are present in at least two of the last three months.

  • At least 50% of five or more recognized fixes represent fulfillment, information, or a change.

A matching new ticket is converted into a service request only when at least 80% of the recognized fixes represent fulfillment. Otherwise, the dispatcher receives a suggestion.