Justifying the need for an APM manager
An Application Performance Management (APM) tool-set is indispensable to ensure superior end-user experience or first impressions of your mission-critical web applications and mobile touch-points. Over the years, APM has transformed into a must-have solution that watches every transaction under the application hoods like an eagle watches its prey sitting high above the cliff of a mountain. Various customers and users of APM solutions have reported a ‘system down’ situation several times a year for maintenance purposes or component corruption reasons. A resolution is met only after issuing support tickets with the APM vendor who utilizes costly support resources to remediate and revive APM. In the process of troubleshooting, customer resources are requested for an assortment of descriptive logs that exhibit errors and problem symptoms. Much time is lost and often customer resources are left with bitter support experiences. Rightfully speaking, your APM gizmo monitors your applications, but who monitors your APM? Enter APM Command Center! During the downtime, several app issues go un-noticed and customer experience suffers. In almost all such cases, issues or problems could have been easily resolved using an APM management service. There is a definite need for an APM manager that manages and administers the performance of the APM solution itself and ensures SLA of core APM services to avert an ‘Application Mayday’ scenario. This is an impressive way of meeting SLAs for your APM solution.
‘Monitor the monitor’ example and inevitable characteristics.
Before we delve into the inevitable characteristics of an ‘APM Manager’ component, let us skeleton the high level functions of this APM service. In order to shortlist the functions, let us use a relevant example from the aviation industry. Almost all modern aircrafts include an Electronic centralized aircraft monitor (ECAM) system that ensures the performance and function of the flight instrument system. In fact, in Boeing A380, a backup fail-safe monitor system manages the ECAM. The key functions of the flight systems monitor are diagnostics, configuration and health management of all on-board flight systems. Similarly, from an APM administration point of view, APM users find it difficult to manage their gargantuan APM infrastructure. The number one challenge with the APM infrastructure is the agents, which could range from a few hundred to tens of thousands in install quantities. The important function required to manage the agents is hot agent configuration and deployment. This could be in the form of cloud based stores (where agent packages are published) to create tailored packages for deployment to application systems. It sounds demanding that self-healing (through process automation) be availed for defective services (core components) of APM, however, such intelligent APM manager component is a crucial component of a true next-generation APM solution.
Figure 1: A simple diagram depicting APM as the application health orchestrator
The real user benefits of an APM manager
An example of an APM manager component is the CA APM Command Center component that enables easy management of the APM solution at scale. Such capability should not only simplify the administration of APM installation but also assist with diagnostics and troubleshooting of defective deployments of agents (monitoring pockets), and bundling and packaging of agent installs so that you can pick & choose, disable and enable monitoring modules for individual applications. To detail the troubleshooting function, would be remarkable to enable remote collection of diagnostics logs from thousands of app machines, so that a central console is used to share information with tool support teams. Just like the core element of modern apps, the APM monitoring management solution should incorporate an API mechanism that assists in quick and bulk maintenance of APM pieces. A well-carved APM monitoring and administration solution must include an agent deployment capability that eliminates the need to hop onto every specific application, database and web server that require monitoring.
This is music to the ears of the APM administrator and application team members that manage the beast (APM), wired across their infrastructure like an orchestrator that run the monitoring musicals as shown in figure 1.