Application performance monitoring (APM) has become a must-have technology for IT organizations. In today’s era of digital transformation, distributed computing and cloud-native services, APM tools enable IT organizations to measure the real experience of users, trace business transactions to identify slowdowns and deliver the code-level visibility needed for optimizing the performance of applications.
2018 will see the requirements and expectations from APM solutions increase in the following ways:
Application and infrastructure monitoring will need to converge
Today, application performance monitoring tools focus mainly on code-level performance. These tools are effective when performance bottlenecks are in the application code (e.g., inefficient method call, poorly designed database queries, external web service causing a slowdown, etc.). At the same time, the performance of applications also depends on the IT infrastructure they operate on. Performance bottlenecks in any of the infrastructure tiers, such as database servers, the virtualization platform supporting the application, the storage tier, or infrastructure services – such as Active Directory, file access etc. – result in application slowdowns. One example of this would be how a storage bottleneck can cause database queries to be slow, which in turn affects the user experience.
Until now, IT organizations have been using different tools for application and infrastructure monitoring. This forces performance diagnosis across application and infrastructure tiers to be performed manually. Not only is this process slow and time consuming, but it also requires domain experts to be involved for troubleshooting.
Prediction: Application managers and IT teams will realize that they need contextual visibility into how an infrastructure problem affects application performance. 2018 will see an increasing demand for converged monitoring tools that include both application and infrastructure monitoring capabilities. These tools will be expected to cross-correlate across application and infrastructure tiers automatically to pinpoint the cause of application slowness: is it in the application code, or is it due to one of the infrastructure tiers, and why?
Going beyond user experience monitoring will be important
Today, there is a lot of emphasis on user experience monitoring. IT organizations are deploying synthetic and real user experience monitoring tools to measure user experience and report on compliance to SLAs.
While it is undoubtedly important, user experience monitoring only focuses on the interactive portion of a web application. In many applications, the processing required to fulfill a user request happens in the background, asynchronously. Failure or slowness of the non-interactive processing tasks affect the web application performance even more than that of the interactive tasks.
Prediction: IT organizations will demand that monitoring tools be capable of monitoring the non-interactive processing tasks associated with their web applications. Since there may not be common interfaces for monitoring such non-interactive tasks, monitoring tools will need to be extensible so IT administrators can add custom monitoring for their applications.
Domain expertise will become necessary
Increasingly, APM tools will be measured on the basis of the return on investment they offer. When a problem occurs, organizations will analyze if the APM tool alerted them to it and whether it pin-pointed the cause of the problem. The use of machine learning, pattern recognition, auto-correlation and other analytics capabilities are important, however, without the right metrics being collected, even the most intelligent monitoring tool may not be successful.
Most APM tools today focus on application code level visibility alone. The application is instrumented for monitoring in a way that is transparent to the application – e.g., through byte code instrumentation. The metrics collected are not application-specific. But for diagnosis to be accurate, application-specific monitoring is necessary. For example, an SAP application may experience slowness because the limit of work processes configured at the SAP server level has been reached. Likewise, SharePoint slowness can occur because growth of the content database is slowing query processing. Application-specific insights are necessary for accurate and timely problem diagnosis.
Prediction: 2018 will see application owners demand more application domain-specific KPIs from monitoring solutions. This will force APM vendors to embed domain expertise into their tools to help administrators quickly and easily troubleshoot problems with business-critical packaged applications such as SAP, SharePoint, PeopleSoft, Siebel and so on.
APM must focus on reporting
Until now, APM tools have been mainly used for real-time problem analysis and diagnosis. IT teams analyze user experience from different locations to point to network problems that affect application performance. They analyze application request flow topology graphs to highlight the time spent processing a request at each tier and to highlight the problematic tier.
But historical performance analytics have not received much attention. APM tools collect volumes of data. The addition of infrastructure monitoring to these tools will only increase the amount of data stored further. IT organizations will realize that operational analytics based on analysis of historical data can be extremely valuable – for application and infrastructure optimization, right-sizing, capacity planning and so on.
Prediction: In 2018, IT organizations will demand that monitoring tools provide built-in capabilities that provide actionable insights based on analysis of the collected data. Prescriptive analytics that spell out actions that IT organizations need to take to get more out of their IT and application investments will be required. APM tools will be expected to incorporate predictive capabilities to forewarn of impending problems. Providing this analysis in a simple to understand format, so even non-experts can interpret and act on the advice provided, will be important.