Monitoring is the Key to Improving Database Performance

Performance is a term that gets thrown around in many different contexts. It’s a much sought-after commodity in endeavors as diverse as sports, automobiles, business, stock portfolios, and IT systems like databases. We are going to take a closer look at what defines performance and how it is measured. From that point, steps can be taken to improve performance in virtually any context.

What is Performance?

Performance can be defined in a variety of ways. One that pertains to our discussion today defines performance as the accomplishment of a given task against preset and known standards. This is a rather abstract definition that can apply to any task with standards that have been agreed upon by the parties involved or imposed by industry or regulatory entities.

Partly due to its widespread use in various settings, the concept of performance can be somewhat nebulous. Logically, it can’t mean exactly the same thing when used to describe entities as varied as athletes, corporations, and SQL Servers. But there are similarities in gauging performance that are applicable in just about any circumstance. Here are some examples of how performance is measured in some very different environments.

  • Sports - The world of sports touches a large segment of the population either as participants or spectators. Gauging performance is usually fairly straightforward in this realm. Team winning percentage, batting average, and point per game are all pretty obvious ways to measure the success of a team or player. But sheer numbers are not always a complete indicator of a player’s performance. A baseball pitcher may have a losing record due to the efforts of their teammates despite limiting the opposition to a single run per game. So when sports are involved, performance needs to be looked at with a discriminating eye.

  • Stocks - Many individuals have a financial stake in the performance of the stock market either through direct investments or a retirement plan such as a 401K. When selecting the stocks to purchase with their hard-earned money, investors have some specific metrics that can help guide their decisions. Some of the most important include return on investment (ROI), earnings per share (EPS), and return on equity (ROE).

  • SQL Servers - The popularity of the SQL Server database platform makes it very likely that you will be tasked with maintaining system performance on an instance if you are a database administrator. As with the other examples discussed, some key metrics can be used to measure the performance of your SQL Server databases. Among them are row counts, file input and output, wait states, and transaction log size. Experienced DBAs should have other statistics that they have learned to trust when investigating performance issues.

Monitoring is Like Keeping Score

A reliable method of measuring performance is to monitor metrics regarding the entity in question against a baseline. In some cases, it is pretty obvious what needs to be tracked and measured. In other situations, the defining characteristics of high-level performance can be more obscure and hard to pin down. For this reason, it’s better to observe multiple aspects of the given task or entity when creating a baseline that can be used for comparison and determining whether performance is increasing, decreasing, or remaining static.

An apt analogy can be made between a DBA monitoring SQL Server metrics and the general manager of a major league baseball team. Concentrating on a single metric to the exclusion of other factors can give a warped and incomplete view of the system or player being reviewed. This can result in bad decisions that have far-reaching impacts on an organization.

A good general manager takes into account many aspects of a player’s performance before offering them a contract. Identifying how recent performance stacks up against previous seasons enables the executives to make educated decisions regarding the potential value players can provide the team. Whether or not to re-sign a player and the terms of the contract are determined by comparing current performance against the baseline of their career stats and those of the wider population of comparable professionals.

Likewise, a DBA needs to monitor a wide variety of metrics to keep their SQL Servers performing at a high level. Focusing on a single issue such as disk space may be necessary to address a specific situation but cannot be the only area that receives attention. Keeping tabs on all the moving parts of an SQL Server implementation that might affect its performance manually is not a realistic expectation. Monitoring tools are needed to assist database teams to maintain system performance.

Monitoring the Performance of Your SQL Servers

IDERA’s SQL Diagnostic Manager for SQL Server can be just what your team needs to optimize the performance of your database environment. It provides a complete picture of your virtual and physical on-premises and cloud SQL Server instances. The tool furnishes DBAs with continuous monitoring and real-time analysis capabilities designed to indicate areas ripe for performance optimization.

Customized alerts can be created to warn when baseline thresholds are being approached or exceeded so proactive action can be taken to avoid performance degradation or outages. Notifications can be sent to specific groups based on the alert’s details to get the right people looking at the problem quickly. In addition to the internal metrics of SQL Server, the application can also monitor OS metrics and alert on the success or failure of jobs.

The features that SQL Diagnostic Manager for SQL Server offers your database team enable them to monitor all aspects of your databases and improve their performance. It’s a valuable addition to your DBAs’ software toolbox.

Anonymous