Machine Learning on Oracle Enterprise Manager Data

Machine learning can be used for a variety of roles such as troubleshooting, capacity planning, and performance analysis.

Businesses can use machine learning to understand seasonality for trending and forecasting capabilities or display dashboards on data already collected.

During a DBTA webinar recently, Courtney Llamas, Architect, Strategic Customer Program Systems Management and Security, Oracle; and Timothy Mooney, Director, Oracle Product Marketing, discussed taking a proactive, top-down approach to SQL tuning by leveraging Oracle’s advanced analytics to see which databases, and specifically which SQL IDs, are seeing their performance degrades over time.

Common tasks in Oracle Enterprise Manager for a DBA include alerting, tracking historical metrics, performance tuning of a particular database or SQL, database management, etc., Llamas and Mooney said.

The types of data collected consist of metadata, metrics, events/availability, and configuration.

However, database administrators face a multitude of challenges, in various areas:

Resource utilization optimization

  • What resource will become the system bottle based on the current growth in the load of the embedded databases?
  • Which systems have spare capacity to support additional database consolidation?

SQL performance optimization

  • Which SQLs work better or worse than expected?
  • Which of the bad SQLs regressed due to bad plans or excessive expectations?
  • Is application performance suffering due to suboptimal SQL?

Troubleshoot issues

  • Which databases are reporting the most issues and what type?
  • How do I correlate application level and database to perform RCA?

Oracle Management Cloud gives users the ability to troubleshoot some of these issues and answer these questions, Llamas and Mooney said.

OMC features include:

  • Global Threat Feeds
  • Cloud Access Identity
  • Real users
  • Synthetic Users
  • Application Metrics
  • Transaction Server Metrics
  • Diagnostic logs
  • Host Metrics
  • Virtual machine metrics
  • Container Metrics
  • Configuration compliance
  • Tickets and alerts
  • Security and network events

An archived on-demand replay of this webinar is available here.