SQL Server is an excellent platform to use as a backend for various database applications. It offers a fast and stable basis that allows the application to run fast, increasing the end user productivityand faster turnaround for a company. Coupled with a reliable tool with a graphical interface such as SQL Server Management Studio it makes creating tables, data manipulation, functions and stored procedures development, resources optimization, etc., to be performed in no time at all.
In the beginning, applications are created and optimized in a development environment, fine-tuned in a test environment, and that allows the application to run fast in a production… at least at the beginning.
Over the time as use of the application increases as well as the amount of data that is stored in the database, manipulation with data starts to be slower that causes that application starts to suffer from some performance degradation which inevitably leads to end-user complaints. The actions they performed without waiting now takes several seconds or even minutes, reports that are generated in a couple of minutes now takes an hour or two
And this is where various techniques for SQL Server performance monitoring and tuning come into play. SQL Server performance monitoring tools should be used not only as a tool for resolving the issues that are already arisen but also in a regular SQL Server maintenance and for supporting and troubleshooting the performance problems at their early stage or even before they arise and thus keep the applications and SQL Server running at peak performance
This article will deal with the top things that SQL Server performance monitoring tool should have to be able to help database administrator in keeping the SQL Server in a top condition:
windows and SQL Server performance counters Wait statistics analysis Query waits analysis Information about locking and blocking Query performance info Query execution plan analysisOne of those tools isApexSQL Monitor. ApexSQL Monitor is and web-based SQL Server performance monitoring tool capable of tracking SQL Server and underlying Windows OS performance counters, detailed wait statistic and query wait analysis, real-time alerting, index monitoring and comprehensive reporting, to mention just a some of its features
ApexSQL Monitor is suited for real-time monitoring of Windows OS, SQL Server, and database performance metrics using over 30 counters out of the box and enhanced with the ability for adding any existing Windows OS performance counter or creating own SQL Server and database performance counters. ApexSQL Monitor allows selecting SQL Server instances, databases and hardware devices to be monitored or excluded from monitoring. It also grants a high granularity in selecting metrics to be tracked for each monitored SQL Server, database, and Windows OS. That includes the ability to fine-tune monitoring frequency per each individual metric, to specify the alert threshold values using the predefined or baseline thresholds and alert suppression ability via setting the minimum period in which value must exceed the threshold in order an alert to be raised.
ApexSQL Monitor conveys the performance metrics in real-time graphs and stores the corresponding metrics data in a central repository database for later use.
The System performance page displays Windows OS metrics of the selected instance that hosts the monitored SQL Server. The default set of performance metric consists of:
pages input per second processor queue length free space utilization of processor time network utilization paging file usage pages per second average bytes per reading average bytes per write average disk queue length current disk queue length average read time average write time disk reads per second disk writes per second system availability available megabytes of memory
Periods of 30 minutes, 1 hour, 12 hours, 1 day and 7 days can be displayed in the charts
The SQL Server performance page displays performance metrics on the SQL Server instance level:
Batch request per second Compilations per second Re-compilations per second Longest running transaction Full scans per second Index searches per second Forwarded records per second Free list stalls per second Lazy writes per second Lock requests per second Deadlocks per second User connections Target server memory Total server memory Memory grants pending Buffer cache hit ratio Page life expectancy Page reads per second Page writes per second Page splits per second SQL Server agent status SQL Server availability
The Database performance page displays specific database configuration details and information, and the vital database performance metrics for the selected database:
Number of unused indexes Status changes Total database size Data file size Log file size Log growth Transactions per secondEach of the above charts regardless of the performance metric type can be enlarged to full-screen size for more detailed view
ApexSQL Monitor uses a very detailed AlwaysOn Availability Group dashboard. The application can graphically convey and display a complete Windows Server Failover Cluster (WSFC) and complete AlwaysOn infrastructure that is hosted on that WSFC. The graphically conveyed topography of AlwaysOn availability groups can be displayed not only for monitored AlwaysOn AG but the full structure of AlwaysOn AG servers that are hosted on the same WSFC, thought only the AlwaysOn replicas that are parts of the added AlwaysOn Ag can be performance monitored
Tracking the AlwaysOn specific metrics is available via AlwaysOn metrics page. For more details on performance monitoring, AlwaysOn groups check Monitoring AlwaysOn Availability Groups and Performance monitoring of AlwaysOn Availability Groups Part 2 articles
Also, check the Performance metrics quick guide for quick insight into the available performance metrics
One of the important features of ApexSQL Monitor is its ability to track the Wait statistics of monitored SQL Server. One of the first steps in SQL Server performance analysis and troubleshooting is to check the wait statistic of the monitored SQL Server to be able to understand whether some unexpected waits occurs on the server and what wait stats types are involved in that potential performance issue.