Application Performance Management (APM) software is an essential part of developing and maintaining business critical apps, but not every company uses APM tools in the same way. Some use it extensively throughout software development and quality control, while others bring out APM tools to help with production issues that operation managers just can’t seem to identify. Once you’ve persuaded management that APM has a strong return on investment, these are a few key questions that have to be addressed before you begin your selection process and implementation.
Point Solution or End-to-End Monitoring?
APM tools can either dig deep into one facet of infrastructure, or cover an entire application layer to look for problems. Point solutions may look at the network, the server pool, a single database, or even a subset of program logs that may not be operating as intended. An end-to-end solution is better suited for application specific layers including the application repository, application server, workflow analysis, and integration services. End-to-end APM tools provide real-time, contextual insight that operation managers can use to find the source of errors, crashes, and other performance issues.
Not only are these approaches not mutually exclusive, they serve complementary functions that the most successful IT departments use like a one-two punch. First, end-to-end APM tools give a 360 degree view of what’s happening and identify the source of the problem. Then point solutions drill down to provide deep insight into each component identified as contributing to performance issues. Together, these techniques form a powerful approach to business process evaluations. On that note, a recent infographic titled, The State of APM, highlights how end-to-end Application Performance Management is becoming the gold standard for IT organizatons.
Agentless or Agent-rich Architecture?
APM tools can be deployed either as a monitoring agent on servers and desktops (agent-rich), or as agentless monitoring technology that remotely connects to servers and desktops to gather information for user experience monitoring, troubleshooting, and root cause analysis.
In the past, when bandwidth was more expensive and remote monitoring could slow down processes, the decision between these two approaches was more complicated. Now most IT departments prefer the agentless approach because it provides more operational flexibility with essentially no drawbacks. Agentless architecture means that changing the monitoring solution doesn’t affect the production server, making it easier and faster to deploy. It also creates introduces fewer dependencies into the system and reduces the overall complexity of software development and quality control projects.
Unless there is a specific reason why your company prefers and agent-rich APM architecture, going agentless is more convenient.
Plug-and-Play or Generic?
Plug-and-play APM tools have already been tailored to deal with a specific application environment (Oracle Siebel CRM and SAP ERP are common examples), but a generic tool can be configured to monitor anything. If you are using software that was developed entirely in-house, you may have no option but to use generic APM tools, but if you are using a major application suite you will save a lot of time and money with plug-and-play solutions. These APM tools can interpret app-specific log files, evaluate user roles, analyze configuration settings, and handle app-specific tracing information not contained in log files all without manual configuration.
Plug-and-play APM tools are ready to use in just a few days, while generic solutions may take weeks to get up and running, and months of tuning and extra configuration to detects all issues that are critical to your application. Because of the amount of effort needed to keep them working, generic APM solutions may even force you to dedicate staff to configuring them, while plug-and-play tools can even be used by non-experts. Whenever it’s available, plug-and-play APM is the better option.
Of course you may not be able to monitor all business processes the same way. Large corporations especially may have complex application environments that are maintained by multiple teams using slightly different approaches. What’s important is to determine how APM fits into your overall strategy before you get it up and running.