AIOps Release 2.2

Overview

AIOps 2.2 introduces new features leveraging AI for auto resolution of tickets and similarity checks in correlation rules. The release also focusses on redefining the alert and ticket lifecycle in AIOps while introducing flexibility to customize the labels for alert and ticket states.

This release also introduces enhancements in ticket templates, rule operators and ITSM integration configuration for faster customer onboarding.

AIOps 2.2 introduces AI enabled automatic resolution of tickets. This is powered by SmartOps Conversation Designer, Rule Engine and Automation Story. SmartOps Conversation Designer’s NER capability is used to detect the intent of a ticket and to extract entities required for resolving the ticket. Automation Story capability is leveraged to run an automated workflow for resolving the ticket with the details extracted from the ticket as inputs.

A new tab called Automation will be available in AIOps for setting up Auto Resolutions. All the intents created in SmartOps conversation designer as well as the Automation Stories created for AIOps (tagged with #autoresolve) in SmartOps will be listed. An admin can set-up an Auto Resolution by mapping an intent to an Automation Story.

While creating an Auto Resolution, the user defined inputs required for the Automation Story must be mapped to the entities extracted from the ticket. An admin should also configure the minimum confidence score for intent detection for the Auto Resolution to be triggered.

Users can also view details like average confidence scores for intents and total number of intent detections for analysis.

Users can now use ‘similar’ as an option in similarity criteria of correlation rules. This uses an AI based similarity check to determine semantic similarity between sentences. With this addition, three options are currently available in similarity criteria of correlation rules.

• Identical - looks for exact match between values of the selected alert attribute
• Empty - looks for values which are blank, for selected alert attribute
• Similar by N% - looks for N% of semantic similarity between values of selected alert attribute

For example, the following alert messages will be treated as similar by using the ‘similar by 80%’ option in similarity criteria. Using this option will allow more alerts to be correlated into a cluster effectively.

Since this option looks for semantic similarity; it works best when used for larger text which also has context – for example alert description or alert message. For alert attributes like resource name, alert source, etc. which do not show a semantic relation, it is recommended to use operators like CONTAINS, IDENTICAL, EMPTY etc.

AIOps 2.2 introduces defined states and next possible actions, for both alert clusters and tickets. Each state and action can be given a custom label by an ITOps admin. This allows flexibility to customize the state and action labels for different customers.

For example, the state when an alert cluster is ticketed and assigned to a virtual engineer can be labeled as ‘Open’ for one customer and ‘Active-VE’ for another customer.

Default configured labels

The default lifecycle for alert cluster states and ticket states in AIOps will be defined as below, using label configuration. The default states for alert clusters will be - OPEN, ACKNOWLEDGED, ACTIVE, ON HOLD and RESOLVED. The default states for tickets will be - OPEN, ACTIVE, ON HOLD and RESOLVED.

The actions available from each state is also defined. Below are the next possible actions from the default alert and ticket states. The labels for actions can be configured by an AIOps admin.

Two new operators ‘CONTAINS IN’ and ‘EXACT IN’ will be available in the list of operators for rule and policy creation. Both these operators will accept multiple values which are comma separated.

CONTAINS IN will look for presence of anyone among the comma separated values, in the attribute.

EXACT IN will look for exact match of the attribute value with any one among the comma separated values.

From this release, all updates for time field in ITSM will be made in the time zone specified in the Project Specific Time Zone setting. UTC is the default time zone unless otherwise specified in the project settings. Please note that the change is applied for the fields in ITSM corresponding to time data and any time information contained in any other field (like alert description) sill continues to be in UTC.

In this release, ticket templates are introduced in one more ticket update scenario - ‘add new correlated alert into ticketed alert cluster’. With this, admins can now customize the ticket updates in this scenario using ticket template. This is an optional template and needs to be set up only if customization is required.

Also, the options available to AIOps admins is expanded, while using ticket templates. Admins can now refer to base alert, ticketing alert or alert cluster attributes in ticket templates. This feature is available in templates for ticket create and ticket correlate scenarios. For example, during ticket creation, the ticket impact field in ITSM can be updated based on the severity of the alert cluster.

• (5377) Correlation error for an alert in production environment
• (7124) Flapping closure is not working as expected

Release Artifacts

Deployment Artifacts

For support and more enquiries, you can write a mail to us on feedback.smartops@ust-global.com