need to monitor my new application to satisfy my 12 operational principles. In my last post I summarized the different sides of an application that need to be monitored, monitoring tools, and monitoring companies. From my point of view as a sysadmin, that’s the kind of monitoring that’s important. As much as it pains me to say it, monitoring is not all about technology. If an enterprise relies only on monitoring applications, it’s going to get in trouble.
I did not take a holistic approach to monitoring all business components - I did not even try to comprehensively monitor all the technical nuts and bolts - all the networks, hardware, and software. I am only looking at the monitoring of one new cloud-based application. I’m focusing on an enterprise application that cannot enter an operational state until it is monitored.
Application monitoring is a few links in the chain.
Monitoring the hardware and software has nothing to do with IT costs, assessment decisions, or warehouse inventory. I can exclude all those from the enterprise application monitoring requirements. The managers may have been arrested, the building may be on fire, and the customers may have jumped ship — that’s fine, as long as the application monitoring is giving me a green light.
A business administrator must monitor the many moving parts in an internal business environment. It’s the only way to provide excellent customer support. From an enterprise point of view, an application is just one thing in the monitoring landscape. Business areas contain technical areas, and technical areas contain applications.
No matter how business critical a new application is, it is still only one part of the IT estate - and the IT estate is only one part of the enterprise. Monitoring also covers human resources, building assets, business processes, legal contracts and everything else that keep the customer sales flowing. They are all related, they are all necessary, and all need to be monitored.
The best monitoring is linked to business needs.
Application monitoring covers the application itself and the environment it resides in: the system resources making up the technology stack, such as network bandwidth, CPU usage, and disk space. A monitoring application constantly measures components and raises the alarm if the numbers are outside tolerances.
This is great for reacting to problems, but not so good from a business point of view. Gartner defined an IT Infrastructure and Operations Maturity Model with six levels of increasing business value. The starting point is the business hell of level 0 where there is no formal strategy. Business heaven is level 5, where the infrastructure adapts to constant business innovation in real time (many cloud suppliers believe they can help an enterprise reach level 5). In monitoring terms, level 0 is no useful monitoring and level 5 is monitoring everywhere, where the measurement data is turned into business information on-demand.
Ideally I want a monitor that measures how helpful the new application is to the enterprise. Unfortunately, monitoring can help with business needs to an extent, but there are limits. The measurements coming from system monitoring are about mechanical parts, not about business needs.
- Monitoring can’t tell me if my isolated data stores are inefficient, but it can tell me when data sourcing is unacceptably slow.
- Monitoring can’t tell me my responsive design is not rendering correctly on new mobile devices, but it can report on which mobile devices are being used.
- Monitoring can’t tell me if my change control process is helping or hindering management, but it can highlight unusual levels of change requests in the system.
Monitoring an application and the technical platform it rests on is essential for a mature enterprise, but it is only one piece of the monitoring puzzle. All business areas work together to provide a customer service, and all must be monitored. It’s not IT, and it can’t be automated, but unfortunately it can’t be avoided. Comprehensive monitoring enables comprehensive support.