AIOps: Monitoring and Observability - Group home

OMEGAMON Storage dataset attribute groups tips and tricks


The OMEGAMON for Storage dataset attributes groups function allows users to filter the information in the complete Dataset Attribute Database.  A dataset attribute group is the set of datasets that conform to the criteria specified in each dataset attribute group definition. Here are some tips and tricks which can help users to create dataset attributes groups that are efficient and reduce CPU time consumption.

1. When a Dataset Attribute Database collection cycle begins, the information associated with each group is cleared and is updated upon completion of the collection cycle (Full or Incremental collections). Thus, reducing the number of groups and increasing incremental collection interval could help you to reduce CPU consumption for Dataset Attribute Database updates.
PTF UJ06979 for APAR OA62263 allows you to disable collection for any dataset attribute group.  This allows you to update the dataset attribute group only when you need the information.  This works well with dataset groups that are not normally referenced.  You can turn on or off collection for a group as shown below.  

2. Realtime Dataset Metrics (RDM) could be used for reviewing datasets without creating any groups. If needed, the user can save the filter in RDM and it will be appeared in Tivoli Enterprise Portal (TEP) or E3270UI as a dataset attribute group with the prefix “RDM_”.

3. Situations could be specified for dataset attribute groups (attribute group “S3 DSN Attr Group Summary”) or for datasets covered by these groups (attribute group “S3 DSN Attr Group Detail”). The attribute item “Group Name” could be used to run a situation for a specific group with some additional condition. For example, alert when there are datasets in the group, automatically delete datasets in the group or reblock datasets that have a small blocksize value.

4. There are 2 tabs that appear during dataset attributes group creation: “Properties” and “Attributes”. Datasets that match any of the properties and each of the attributes that you specify will be included in the defined group. For properties, the logical OR is used. For attributes, the logical AND is used. Logical AND is used between properties and attributes. As a result, the dataset attribute group definition looks like the following:

(Property1 OR Property2 OR ... OR PropertyN) AND (Attrbibute1 AND Attrbibute2 AND ... AND AttrbibuteM)

Defining a dataset attribute group with multiple properties can be CPU intensive when the group is updated.

5. By default, the first property (or the first attribute if no attributes) is used as the primary index and is used for initial filtering. All other attributes are checked against datasets passed the initial filtering. As a result, the first attribute should filter out most datasets to reduce CPU consumption required for group updates.
For example, you have 10000 datasets. You specify 2 attributes: the first attribute filters 9000 datasets and the second one is only 100 datasets. To get a combination of 2 attributes a second filter will be applied against 9000 datasets from the first filter. In this case, if you change the order of these filters you will have only 100 checks for additional filter and this combination will work faster and consume less CPU time.

#OMEGAMON #IBMZ #IBMZOS #monitoring #AIOps #Storage