AIOps

The Getting started with Watson for AIOps Event Manager blog mini-series will cover deployment, configuration, and set-up of Event Manager system to get you off to a fast start, and help you to get quick value from your investment.

This sixth module considers a second method of doing event correlation: scope-based event grouping. Scope-based event grouping is based on a basic premise that events occurring at the "same place", and at the "same time" are likely related.

For example, if a remote cell site has a power failure, there will likely be alarms generated from equipment there, along with building management alarms, and so on, in a short space of time. In this case, the scope would be the remote location, and the time window might be, say, a 10 minute window. Another example is where there is a hardware failure in one of the servers that contribute to the provision of a service, such as an application or website. This failure may result in alarms being generated from the dependent parts. In this case, the scope would be the common application or service name, and the scope might be, say, a 5 minute window.

Scope-based grouping is provided by Watson for AIOps Event Manager so that operations teams can input local knowledge on how events should be grouped. A UI is provided so that scope-based event grouping policies can be created and maintained.

In practice, scope-based event grouping is very powerful and yields tremendous event reduction by virtue of grouping events, thereby effectively reducing the number of rows presented to operators. The level of reduction achieved varies depending on the data sets, but it is typically upwards of 60%. A major north-American financial institution reported that scope-based event grouping helped them to reduce the number of tickets they open by around 70%, by opening tickets based on groups of events rather than individual events.

By the end of this module, you will have created two scope-based event grouping policies, injected some test events and observed the grouping. The second example will include overlap with a topology-based event correlation scenario in the previous module to demonstrate so-called "super grouping".

This module should take you about 30 minutes to complete and includes the following steps:

Step 1: Create a scope-based event grouping policy to group by Service (5 minutes)
Step 2: Create a right-click tool to inject some sample Service events (10 minutes)
Step 3: Create a scope-based event grouping policy to group by Location (5 minutes)
Step 4: Create a right-click tool to inject some sample Location events (10 minutes)

Step 1: Create a scope-based event grouping policy to group by Service

Log in to Watson for AIOps Event Manager and navigate to: Netcool WebGUI.

Navigate to: Insights → Scope Based Grouping

Complete the following steps to create a policy based on common application name:

• Click on: Create policy
• Give the policy a name - eg: GroupByService
• Complete the Description field with something appropriate - eg: Policy to group events by common service name.
• Leave the Priority field set to 1.
• The default Datasource should already be checked.
• From the Events section, select: Specify conditions
• Set Condition 1 to: Location field Is not empty
• Set the Action parameter to define the scope: set it to Service
• Set the Time Window to 300 seconds (5 minutes)
• Set the Enabled toggle switch to On and Save.

You should end up with something like the following:

Step 2: Create a right-click tool to inject some sample Service events

Use the following SQL to create a new right-click tool to inject some events to match the scope-based grouping policy just created. You can use the same instructions given in step 4 of the previous module to achieve this task.

insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Service) values ('db2server1clean1', 'db2server1', 'Running cleanup automation script', 2, 13, 'Syslog', 'WebService');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Service) values ('db2server1clean2', 'db2server1', 'Error: unable to run cleanup automation script', 3, 1, 'Syslog', 'WebService');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Service) values ('db2server1syslog', 'db2server1', '/dev/sda1 is 100% full', 5, 1, 'Syslog', 'WebService');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Service) values ('db2server1db', 'db2server1', 'Database is inaccessible', 5, 1, 'ITM6', 'WebService');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Service) values ('websphere01', 'websphere01', 'websphere01 unable to access transaction database', 4, 1, 'ITM6', 'WebService');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Service) values ('websphere02', 'websphere02', 'websphere02 unable to access transaction database', 4, 1, 'ITM6', 'WebService');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Service) values ('widgetcom', 'www.widget.com', 'www.widget.com: response time >30s', 5, 1, 'ITM6', 'WebService');

Select any event and run the tool from the Alert Viewer. You should end up with something like the following:

Step 3: Create a scope-based event grouping policy to group by Location

Log in to Watson for AIOps Event Manager and navigate to: Netcool WebGUI.

Navigate to: Insights → Scope Based Grouping

Complete the following steps to create a policy based on common geographic location:

• Click on: Create policy
• Give the policy a name - eg: GroupByLocation
• Complete the Description field with something appropriate - eg: Policy to group events by common location.
• Leave the Priority field set to 1.
• The default Datasource should already be checked.
• From the Events section, select: Specify conditions
• Set Condition 1 to: Location field Is not empty
• Set the Action parameter to define the scope: set it to Location
• Set the Time Window to 300 seconds (5 minutes)
• Set the Enabled toggle switch to On and Save.

You should end up with something like the following:


NOTE: The Priority specifies the order in which the Policies will be tried. Once a scope-based event grouping Policy matches and the scope is set for an event, all subsequent Policies will be skipped. If two policies have the same Priority value (as in this case), the execution order is indeterminate.

Step 4: Create a right-click tool to inject some sample Location events

Use the following SQL to create a new right-click tool to inject some events to match the scope-based grouping policy just created. You can use the same instructions given in step 4 of the previous module to achieve this task.

insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Location) values ('rackserver01', 'racksvr01', 'Warning: hardware fault detected in processing array', 5, 1, 'SNMP', 'London01');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Location) values ('esx1', 'esx01', 'Warning: low resources', 3, 1, 'VMware', 'London01');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Location) values ('computer1000esx', 'computer1000', 'Auto-scaling failed: unable to allocate resources', 4, 1, 'VMware', 'London01');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Location) values ('computer1001esx', 'computer1001', 'Auto-scaling failed: unable to allocate resources', 4, 1, 'VMware', 'London01');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Location) values ('server100esx', 'server100', 'Auto-scaling failed: unable to allocate resources', 4, 1, 'VMware', 'London01');
insert into alerts.status (Identifier, Node, Summary, Severity, Type, AlertGroup, Location) values ('server200esx', 'server200', 'Auto-scaling failed: unable to allocate resources', 4, 1, 'VMware', 'London01');

Manually delete the events created in the previous module and re-run the tool that created these events along with your newly created tool. You will observe three groups of events merged together: two created by topology-based event grouping, and one created by scope-based event grouping.

NOTE: You may need to wait a few seconds and refresh the event view before the grouping is visible.

Select any event and run the tool from the Alert Viewer. You should end up with something like the following:


You can click on the grouping buttons on the right to inspect the details of the group. Topology-based grouping will show a preview of the topology that has effected the grouping. Note how you have events grouped together by both topology-based grouping and scope-based grouping in the "super group". The events from computer1000 and computer1001 are found in both the topology-based correlation template groups, as well as the scope-based event grouping policy. Hence these events are members of two different group types, and so the two groups are automatically merged together due to these common events.
--
You have now completed this module and are ready for the final module in this series, module 7: Enable optional grouping features.
#CloudPakforAIOps
#EventGrouping
#NOI
#HowTo