We all rely on services every day that are delivered through integrated or connected IT systems.
Services that are easy to identify include: accessing transportation time tables through an app, purchasing transportation tickets online or through a kiosk, using your loyalty program in a local store or retrieving your medicine at a pharmacy. Behind each of these services there are several software integrated or connected to deliver the service to a user. These services often rely on complex logic and communicate in real-time to give you a real time update of the time table, the right pricing of merchandise you buy and the right medicine.
The number of services we are exposed that are delivered through interconnected IT systems grow quickly as a consequence of the digitization race, explosion of cloud applications and use of mobile devices /apps.
Monitoring has traditionally focused on infrastructure. Is the CPU or Memory load under a certain threshold and is my host running. That’s antiquated. As end user of a services provided by IT infrastructure we do not care about the availability of the infrastructure – we care about availability and response time of the service we use. That means that businesses and public services needs to measure the performance of the applications or IT systems that are integrated to deliver the services to the users.
The antiquated solution to deliver services built on multiple software applications or IT systems was to develop custom one-2-one integrations between applications – or spaghetti wiring / integration. Spaghetti integration is costly to implement and maintain and is hence acknowledged as a poor solution to the integration need. Integration middleware is the alternative for many or using modern microservices architectures.
Independent of your integration architecture, keeping control of the performance of the integrations between the applications will be core to monitoring performance of the services end-2-end. Your integration architecture will likely be the heart of your IT infrastructure that many other services see as the weakest or most vulnerable link that all other systems rely on.
That means you need to ensure you are able to monitor the performance of your integration architecture from a line-of-business (service) perspective that managers and executives care about.
“And don’t even think about telling business owners that the infrastructure is running when millions of sales was just lost due to a performance hiccup.”
So, to provide useful monitoring and insight into the services that are delivered from connected IT systems you need to START OUT with properly monitoring the integration architecture using anomaly detection that identifies performance vs normal behavior. Using Anomaly Detection you establish a baseline for what is normal and avoid the crywolf alerts from manual alerts and you get insight that is relevant for business. THEN add custom anomaly detection from specific systems that are relevant to the services that your integration architecture supports. That allows you to develop end-2-end insight, analytics, anomaly detection and performance monitoring 360 degree and full cycle across your important business services.
Maybe interested in the this recorded webinar?