Documentation

⚲
No results
    gitHub

    Table of contents

    OCI monitoring of Model Hub

    This page provides a step-by-step visual guide for monitoring an Autonomous JSON Database on OCI, based on this Oracle documentation.

    Open console screen for your database

     

    View high-level health in the Monitoring tab

    On the database Details page, click Monitoring (this shows the default charts: CPU, DB Time, logons, storage, I/O, etc.).

    • Use the Interval and Statistic controls on each chart to change granularity.
    • This view is backed by the oci_autonomous_database metrics namespace.  Consult this documentation page for further details.  

     

    Quick reads

    • CPU Utilization (percent across consumer groups): sustained >~80% suggests need for scaling or tuning.
    • DB Time (sec/sec): rises with latency and call volume, if DB Time climbs while CPU is low, wait events/I/O are likely.
    • Current Logons / Sessions: concurrency pressure; combine with CPU and waits.
    • Read/Write Throughput & IOPS: large JSON scans or unindexed filters. Consult this documentation page for further details.  

     

    Tip: Many teams set an alarm on CpuUtilization > 80% for 5 minutes to catch hot spots early. Consult this documentation page for further details.

     

    Deep dive in the Performance Hub

    From the Details page, click Performance Hub (or go via Database Actions → Performance Hub).  This opens the main tuning workspace.  Consult this documentation page for further details. 

    What you’re looking at

    • Time Range / Time Zone selectors at the top control all charts.
    • Activity Summary (Average Active Sessions) breaks load into CPU, User I/O, and Wait — it is great for spotting what you’re constrained by. Consult this documentation page for further details.

     

    Tabs you’ll use most

    • Activity (ASH Analytics): slice active sessions by waits, SQL ID, user, module — use it when DB Time is high but CPU isn’t.  Consult this documentation page for further details. 
    • SQL Monitoring: shows heavy SQL with plans and execution stats; click a SQL ID to open a live report.  Consult this documentation page for further details. 
    • Reports: generate AWR and ASH reports for sharing/offline analysis (Typical/Basic levels).  Consult this documentation page for further details.  

    How to interpret quickly (cheat sheet)

    • CPU-bound (AAS mostly CPU) → add OCPUs (autoscaling/ECPU helps) or tune top SQL. docs.oracle.com
    • I/O-bound (AAS mostly User I/O) → add/filtering indexes (e.g., JSON Search Index), reduce scans, check read/write charts. docs.oracle.com
    • Wait-bound (AAS mostly “Wait”) → use Activity/ASH to pinpoint the wait class (locks, concurrency, network). docs.oracle.com

     

    On this page