Data collection steps for validation of new machine learning models

Products

• Mainframe Operational Intelligence

Issue/Introduction

The MOI team is continuously working to improve our machine learning in Mainframe Operational Intelligence (MOI). We are currently validating a new machine learning model--- evaluating improvements in anomaly detection, Green Highway visualizations, performance, and how much time would be required to migrate existing appliance data to the new model.  

• Information about your metric instances, extracted from one of the MOI internal databases
• List of metric instances specifically processed for machine learning  (aka Green Highways and Alerts), including the interval of the metric
• Data from selected MOI logs, e.g. extracted Time Series data (1 month) for a few MOI internal metrics like MOI Appliance → Database Activity → ROWS_WRITTEN

 

Environment

Release : 2.0

Component : MF OPERATIONAL INTELLIGENCE

Resolution

Note:  File   moi-extract-policy.tar_1623854555497.gz    containing the needed components for this data collection is attached to this knowledge document. 

Please follow the steps below to collect the requested data:

1. On the MOI appliance, create a new folder under the /tmp folder:  mkdir -p /tmp/moi-extract-policy

2. Download the attached moi-extract-policy.tar_1623854555497.gz from the support case and place it under /tmp/moi-extract-policy/

3. Extract the package with this command:   tar -xzf moi-extract-policy.tar_1623854555497.gz          .  

4. Start the data extraction with this command:  sh extract-policy.sh

5. The utility will run approximately 5 mins. Upon successful completion, the script will generate file  /var/opt/moi/share/moi-data.tar.gz  

6. Upload /var/opt/moi/share/moi-data.tar.gz to a Broadcom Support case opened for this data collection