A great part of managing IBM Informix Dynamic Server involves performing regular administration and maintenance tasks, responding to unexpected system events and keeping your database server tuned for optimal performance. In the Server Studio JE, the Sentinel service plays several major roles in providing comprehensive performance monitoring, tuning and event response automation for IBM Informix Dynamic Server.
Performance of your database server, to a great degree, depends on the kinds of applications that you run on your system. The causes of performance problems on OLTP (online transaction processing) systems are very different from the causes of problems on systems that are used primarily for DSS (decision support system) query applications. Systems with mixed use provide an even greater performance-tuning challenge and require a more sophisticated analysis of performance-problem causes. Some configuration adjustments can unintentionally decrease performance or cause other negative side effects. Before you begin to adjust the configuration of your database server, it is vital to evaluate the performance history and profile of your current system configuration. Equally, when you do make configuration adjustments to your system, it is important to measure objectively the effects of such changes on the performance of your database server.
Sentinel enables you to monitor in real time resource use of your database server, ongoing transactions and query performance. It also accumulates detailed time-series data on performance history and profile of your system's resource-utilization. For example, you can track the CPU utilization, paging-out rate and the I/O transfer rates for the various disks on your system, to identify peak-use levels, peak-use intervals and heavily loaded resources. Similarly, you can monitor fragment use, to determine whether your fragmentation scheme is correctly configured. If your system's throughput and response times suddenly degrade, you can capture and analyze SQL statements executed by users to determine if particular queries are the cause of inadequate performance. To customize monitoring solution for your system's resource use that are appropriate for your database server configuration and the applications that run on it, the Sentinel gives you the ability to define your own monitors from a matrix of nearly 100 IDS performance parameters.
With detail performance history and profile of your system's resource-utilization information on hand, you can begin to track the cause of problems as soon as users report slow response or inadequate throughput. If history is not available, you must start tracking performance after a problem arises and you might not be able to tell when and how the problem began. Trying to identify problems after the fact significantly delays resolution of performance problems.
Ability to track the cause of performance problems and correct them quickly is good; however, the ability to prevent them from occurring is even more desirable. It is better to be proactive than reactive. Sentinel enables you to create multi-level alerts that notify you of impending performance problems or critical events your database server encounters. Moreover, you can create customized autonomic responses that are triggered by such alerts and execute a series of OS commands, SQL scripts and IDS native utilities. For example, if an 80%-full threshold alert is generated by the dbspace monitor, a custom job script can be automatically triggered to add an additional chunk to it. Regularly scheduled preventive maintenance tasks, such as backing up your databases, can be also automated.
See Also: