SPEEDGAIN FOR DB2
Complete database server monitoring
A complete picture of the DB server is prepared. Often, pure database monitoring is not enough for analysis. Information from the point of view of the operating system complete all the evaluation possibilities.
SQL workload-based analysis
As a basis for optimization in accordance with use, the SQL workload that is available in the database is evaluated and filed optimally for analysis purposes. The user is provided with the means to analyze the aggregated SQL behavior.
Permanent performance monitoring of the server system
Different system resources (e.g., CPU, network, I/O) are continually monitored and objectively assessed using lightweight performance pings. Weaknesses in the system performance can be identified independent of the application load.
In addition, different systems can be compared with each other using objective assessment criteria.
End to end monitoring
Communication between the client and database server is monitored. This means both latency and bandwidth are considered.
From problem to solution
The consequent integration of proven methods of analysis helps the user in effectively solving problems. By a process of elimination, the bottleneck resource is recognized and, using the causal principle, the responsible factor identified.
Easy to install and a resource-saving operation
Easy installation without agents, low system requirements, and minimal application of system resources make for quick set-up.
Intended for in-depth causal analysis and 24/7 operation
Causal analysis is supported optimally by providing an overview of all necessary performance data. In addition, necessary functions, such as alarms and status monitors, are available for 24/7 operation.
Continual assessment of database performance
External database performance from the application point of view is continually assessed and visualized.
The application performance is only influenced when this indicator worsens.
The DBA can quickly assess whether the application is compromised or not with this indicator.
From problem to cause in one click
The exclusion principle:
- Identification of the bottleneck resource
- Which resource is being used with which compartment (CPU, IO, LOG, etc.)?
The trigger principle:
- Which application or SQL is causing which section in the total consumption of the bottleneck to use up resources?
Flexible expansion of monitoring through the module plugin process
New functionalities can be expanded dynamically in Speedgain though the module plugin process. That makes, e.g., integration of your own monitoring scripts possible.
Analyze access paths using Speedgain SQL Workbench SQL
The Speedgain SQL Workbench recognizes and corrects performance problems from individual statements. It can be implemented independent of the Speedgain Repository and is consequently an important means of support in the DB2 application development.
The status monitor shows the current status of all servers and databases on one page
Server operation system information:
- CPU and IO usage directly delivered from operating system
- SWAP area and information about blocked processes
- Load information (logical and physical read access)
- Database time distribution – which resources are using database time
The Speedgain Cockpit offers a quick overview of the database activity and resource usage.
Workload – show activities in the database. Rows read as well as logical and physical reads are displayed, among others.
Application states – shows the status of the applications Distribution – shows the distribution of usage times on the resources such as CPU, I/O time, or sort time.
PDB High Watermark – shows current usage in relation to absolute maximum usage measured.
Identification of "locust swarm"
Identification of allegedly “small” SQL statements that negatively affect the system performance by being carried out frequently.
Switching between the individual SQL statements and an aggregated display is possible in one click.
OS monitoring provides the operating system indicators necessary for an overall picture
Display of correlated operating system and database performance metrics, preparation of server performance indicators for causal analysis
- Number of interrupts and context switches
- CPU load
- Storage usage
- SWAP area usage
- Detailed I/O information
Predefined and configurable reports offer numerous evaluation options
Expanded performance analysis though the use of predefined reports, which can be configured easily:
- Provides insight into hidden and slow developing problem
- Display of resource usage
- Visualization and analysis of performance metrics within configurable time frames
- Growing number of reports
Identify the main users of resources with the ABC analysis
The ABC analysis divides objects into three classes. It is a method of identifying, separating, and prioritizing critical objects.
Division of classes:
Class A – 80% – Class B – 15% Class C – 5%
This enables a focus on the most important triggers and, therefore, enables efficient analysis.
The evaluation of the db2diag.log provides information on performance and availability
With the help of the navigation function, it is possible to check a whole time period, not just a single point, within a list view.
This make interesting log entries, which could be problematic, easier to read and more obvious.
Numerous filters make effective evaluation possible.
Transparency regarding database performance
The performance of all services provided by the database is made transparent for the whole system. This allows assessment of database performance in the context of overall performance.
Improved database service on the basis of historical causal analysis
During a causal analysis in the database area, all necessary historical information from the database server is made available. Database performance data and data from the operating system and the network form part of this process. This enables a retroactive and complete analysis of specific problems and leads to improved database service quality.
Recognition of "locust swarm" through SQL workload analysis
Seemingly harmless SQLs that are executed frequently may affect the performance of a database strongly – similar to thousands of pinpricks or a swarm of locusts. This is precisely because, if they are taken in isolation, they are “small and inconspicuous” and are not perceived in SQL analysis and reports, even though they consume a lot of the total resources.
Through the analysis of the whole SQL workload, this swarm intelligence is made visible and the production of indexes much more efficient.
Early detection of performance blocks
Like the canary in the coal mine, Speedgain detects changes in system performance early on. This enables greater system availability and proactive capacity management.
Is it a database problem?
Through monitoring the produced database performance, it can be quickly detected whether there is a problem in the database or not. Protracted analysis leading to a conclusion of “it must be somewhere else” are superfluous.
Competent help when it's necessary
All support and development is provided from Germany. In a worse-case scenario, there are experienced DBAs at your side who are able to set up Speedgain themselves.
Cost savings though a one-stop solution for performance management of DB2, Oracle, DataStage, and BigSQL
Employees require less time for familiarization, as all Speedgain products are built in the same intuitive way. Usage is more efficient, and the quality of the service is improved.