The Oracle Diagnostics Pack® is a set of applications integrated with Enterprise Manager® that provides the tools required for monitoring the Oracle environment, including: databases, host systems, web servers and Oracle applications.   A part of the Oracle® Enterprise Manager product family, the Oracle Diagnostics Pack covers the key facets of Oracle diagnostics: lights-out monitoring and notification, real-time diagnostics, historical data analysis and web-based reporting.

• The Advanced Events is a set of pre-defined, customizable diagnostic tests that work with the Oracle Agent to provide automated lights-out monitoring, problem notification and resolution for Oracle environments. 

• The Performance Manager is a graphical monitor for performing real-time Oracle diagnostics. Performance Manager includes specialized charts such as the Database Health Overview, which provides a single consolidated view of key database and host performance metrics.  

• The Capacity Planner automates the process of collecting and analyzing historical performance data for the Oracle environment for use in diagnosing problems and planning system requirements. 

Monitor System Activity 24x7
You can�t personally stand watch over your system twenty-four hours a day, seven days a week.  Fortunately, Enterprise Manager provides automated monitoring using the Oracle Agent.   The Oracle Event Management System uses the Agent to autonomously monitor and react to problems detected on managed servers. The Diagnostics Pack provides a set of pre-defined event tests, known as Advanced Events, that use the Agent�s event system to automatically monitor conditions and detect problems for key elements of the Oracle environment. 

The Advanced Events can be configured to check one or more conditions for a target system.  Thresholds and number of occurrences can be customized, as well as event notification plans and corrective actions.   Event thresholds can also be shared by and displayed within the Performance Manager diagnostic charts. The events include hundreds of tests that span key elements of the Oracle environment: 

• Database event tests can monitor database fault conditions, instance performance, space usage, resource contention and other key factors.  

• Host system�s can be monitored for a range of factors including: availability, CPU usage, memory paging and disk space availability.  Extensive node monitoring can also be performed for specific operating systems, including: Sun Solaris, HP-UX, IBM AIX, TRU64 and Windows NT and Intel Solaris 

• Web server�s can be monitored for availability and other activity such as the volume of bytes and requests per second. 

• The optional Management Pack for Oracle Applications adds event tests such as Concurrent Manager locks, request volumes, pending and failed requests.  

• The optional Management Pack for SAP R/3 adds event tests that monitor SAP resources such as Buffers, Memory Roll and Page Areas, and response time.  

The Event system can also be extended to handle custom monitoring requirements.  Administrators can write their own SQL and PL/SQL tests using the User-Defined SQL event feature.  For even greater flexibility, the User-Defined Events feature allows administrators to define events based on their own custom monitoring scripts, using any language supported on the monitored node.   Thus administrators can use Enterprise Manager as their centralized monitoring system,  including monitoring of events specific to their environment. 

Events occurrences can be investigated directly from the console�s event viewer which allows the user to launch real-time diagnostic charts, drill down into historical data and access advise regarding the metric being monitored.  In addition, the Event Handler feature allows administrators to provide custom responses to event occurrences.   For example,  administrators can use the event handler's command execution capabilities to invoke a third party trouble-ticket system that will log a trouble-ticket when an event occurs. 

Troubleshoot Performance Problems
When performance problems emerge, real time diagnostics may be necessary to investigate the source of the problem.  The Performance Manager application provides a real time interface to current system activity that helps you locate and investigate the nature of the problem.   Performance Manager charts can be used to observe high level activity such as host resource usage and instance performance, as well as focused activity, such as response time and resource contention and top resource users. 

Performance Manager charts provide a built-in diagnostics process that alerts the user to problems and trends directly within the diagnostics charts.  If performance thresholds are exceeded the user is notified by a warning or alert flag in the chart.  The user can quickly diagnose the problem using a set of associated drilldowns, including: detail charts, historical data, advise and tuning tools. 

In addition to diagnosing the problem, the user can automate the process of monitoring performance metrics using agent-based events.  Setting an agent event can be done directly from a Performance Manager chart.  The thresholds set for the event can then be shared by the real-time chart to enable consistent performance monitoring policies.

Focus on Key Activity
When it comes to diagnostics, often the most difficult hurdle can be deciding where to start.  Performance Manager solves this problem with the Database Health Overview.  The Overview provides a consolidated view of key database and host machine performance metrics organized into three performance hemispheres: CPU, Memory and Disk, making it easy to see the relationship between host resources and database performance.   The Overview chart also includes sections that highlight top resource consumers and contention activity. 

The Overview chart provides the best vantage point for using the Performance Manager�s diagnostics methodology of threshold alerts, trend indicators and drilldowns.   Using these methods, this single chart provides an entry point for the vast majority of diagnostics monitoring and actions required for an Oracle database. 

The Database Health Overview chart provides a single view of database abnd host performance and includes a built-in diagnostics methodology for drilling 
down to source of the a performance problem. 

Performance Manager also provides special charts that focus on typical troubleshooting scenarios:

• TopSessions, TopSQL and TopObjects allow you to focus on sessions, application activity, or accesses to schema objects that are consuming the most resources. 

• Lock Monitor lets you focus on database locks that are blocking other user activity. 

Plan Future System Requirements
Gathering system usage and performance data over time is an essential part of system diagnostics.  Oracle Enterprise Manager provides an agent-based data collection framework consisting of a low overhead mechanism for collecting performance and system data.  Data can be collected for key elements of the Oracle environment, including: databases, host systems, web servers and applications. 

Data collections are administered using the Capacity Planner application.  Each target type has a set of pre-defined metrics that can be gathered.  In addition, users can add their own custom data collections for supported target types.  Users can select any subset of statistics to collect and the interval at which they should be sampled. 

To control the amount of data that is stored over time, Oracle Diagnostics Pack�s Capacity Planner allows users to establish management policies for the collected data.  Users can specify how often collected data should be loaded into the historical database.  This allows users to schedule the load when network and system resources are available.  Capacity Planner also aggregates and rolls up data automatically.  For example, at the end of each hour, minute data is rolled up to hourly data.  This gives users both a detailed and a high-level view of the system�s performance data.  Users can also set policies for how long data should be stored before it is aged out, thus providing control over how much storage is needed for the historical data.

The data collected by Capacilty Planner can be used to study database system usage and performance and predict resource needs over time.  Analysis charts can be created that include metrics from disparate targets, for example databases and host systems can be included on the same Analysis.  This allows the user to visually correlate related metrics from different sources.  Analysis charts can be easily modified to reflect different date ranges and time granularity such as minutes, hours days and so forth. 

Using the Trend Analysis Wizard, users can forecast when a particular event will occur.  For example, system managers can predict what their disk requirements will be at the end of the year or predict when they will need an additional disk.  Web administrators can determine when they will need to add a new Web server to manage their increasing volume of traffic. 

                              The Capacity Planner Trend Analysis Wizard provides an easy way of analyzing historical data and producing trend analyses. 

Simplify Diagnostics with Web-based Reports 
Reporting is a key method used by Enterprise Manager for providing easy, open access to system management information and diagnostics.   Enterprise Manager provides a robust reporting system that includes many pre-defined diagnostics reports.  These reports can be published to your Enterprise Manager website for browser access by system administrators. 

Default reports are provided for database configuration and status which provide information on database usage and performance.  Reports can be customized by selecting a subset of the reporting information and creating new report templates.  The diagnostics reports can be scheduled for generation on a regular basis or configured to be updated automatically when opened by a user.

The Diagnostics Pack uses the reporting system to make two types of diagnostic information readily available in HTML format: 

• Real-time charts that display current activity and system status. 

• Historical data for evaluating activity and trends over a specific time frame.  

Build Your Enterprise-Wide Diagnostics on a Scaleable Framework

Administering and monitoring large dispersed database environments requires a framework capable of handling hundreds of services and multiple administrators at different locations.  Enterprise Manager provides a scalable framework and advanced functionality built to handle these tasks. 

• Multiple administrators can access and share a common set of diagnostic event tests and data collections using a central Management Server repository.  

• Database administration managers can control which managed targets are available to each administrator through the Enterprise Manager console.  This pevents administrators working in large environments from being overwhelmed by the volume of managed targets and lets them focus on what they are responsible for. 

• Systems that are geographically dispersed, or that span across a mix of LANs and WANs, can be divided into "management regions" and centrally managed by Enterprise Manager.  The Management Regions feature allows the administrator to divide the work of managing a large number of nodes across multiple Oracle Management Servers.   Diagnostic event monitoring and data collections being performed on a node in a Management Region will only be serviced by the Management Servers for that region, thereby improving performance and enabling administrators to effectively monitor geographically disperse targets.  

 

KEY FEATURES
Automated Lights-out Monitoring  Agent-based event tests monitor key performance factors of the Oracle environment  Automated event test notification with shared access across multiple administrators  Immediate access to diagnostics charts, historical data and advise for responding to event tests  Real Time Troubleshooting Graphical performance charts that focus on key performance problems  A single performance overview of host and database activity   NEW!  TopObjects charts to find the "hot" tables, indexes, etc. Diagnostic drilldowns and advise  Fast access to historical data for solving current problems	Web-based Diagnostics HTML-based performance charts for host and database monitoring from your EM website  HTML-based diagnostics advise  HTML-based historical data charts and trend analysis reports automatically published to your EM website  Performance Manager application and special charts such as TopSessions can be run in browser-based thin client mode  Integrated Tuning Tools Integrated Graphical Explain Plan for SQL troubleshooting  SQL Analyze launch in context for fast access to SQL tuning
RELATED PRODUCTS AND SERVICES The Oracle Tuning Pack is integrated with the following Oracle management applications: Oracle Enterprise Manager -- for easy discovery and management of database, nodes, web servers and other targets.  Oracle Intelligent Agent -- for lights out control of diagnostics events  Oracle Tuning Pack -- for launching SQL  tuning tools from SQL diagnostics