Autonomous Database Learning Lounge

By Marcos Arinciba hosting Debasish Sinha

Objectives

• Reduce Cost
• Ease Workflow
• Accelerate Succes

Main Focus on Automatic Issue Detection and Resolution

• 8k+ Metrics used, 1500+ alarms used (alarm result in pager calls to people)
• Monitoring of Load Balancer, Databases etc
• Its all needed because of the scale at which Oracle Operates (for e.g 42 billion queries in 7 days is one such metric)

ADB Operating Model

As of now approx 83% issues are solved automatically.

• Canary is simulating customer env on oracle side to detect failures before it even reaches the customer.

• LCM Operations is another important alarm

• AIM: Autonomous Incident Monitoring for Database Incidents

  • 

• AIC: Autonomous Incident Console

  • 

• AIC: Enhanced Code Drift Analysis

  • AI based workflow to detect potential code transactions based upon symptom (??? Need some explainer, how it works, does it Apply the fix and more).
  • 

• ADC: Automatic Diagnostic Access

  • Automatic Triage is one such system (AI?? Need to Confirm) which processes Diagnostics.
  • 

• AIM: Autonomous Incident Monitoring for Database Incidents

• For Backend Log Monitoring OpenSearch and SQLStats

• Online Patching Change Oracle Binary at Runtime.

  • CI / CD pipeline from Oracle Internal Docs.
  • 

CAPA

The methodology for Oracle to learn from the whole ADM pipeline