They were simpler times a decade ago. IT monitoring focused on the metal. We targeted the servers and switches and monitored network throughput and the CPU spikes. So straightforward.
But all things in IT move quickly, and the days of infrastructure-first monitoring are waning. It’s a whole new approach today.
A history
In the past when we built a monitoring solution, it was designed from the perspective of the IT administrator. Naturally, for anyone overseeing a datacenter, that meant the monitoring was focused on the individual hardware components. Are my servers running? Is there available memory capacity? Are there unexplained network usage spikes?
Major vendors built their flagship performance monitoring products to serve this need, and that’s the direction most organizations have taken until recently: Building monitoring from the ground up. Focused on the infrastructure first.
This approach is excellent at detecting when a server is failing, or when storage is running low. However, it doesn’t always translate into application health, or customer satisfaction, which, when you strip away everything else, is ultimately what infrastructure and IT professionals are tasked to support.
The big realization
The situation began to change with the observation that there is probably a better proxy for monitoring customer experience than our individual infrastructure health. Assuming that is true, IT can deliver more business value by focusing on that better proxy. This is especially true as the infrastructure is increasingly removed from the direct control and visibility of IT as a result of embracing cloud computing. Today, companies use an average of eight cloud providers for their various applications and services.
It’s a shift from infrastructure performance monitoring to application performance monitoring (APM).
When we talk about APM, we’re grouping together a few technologies. One is application instrumentation monitoring, which is a solution embedded in your application. This is well represented by the core offerings of New Relic and AppDynamics. The term has also expanded to include user experience monitoring and application topology visualization. All of these are core to understanding customer experience.
This change in perspective began in earnest almost a decade ago but has seen dramatic increases in speed more recently.
Gartner has noted that “by 2021, 60 percent of IT monitoring investments will include a focus on business-relevant metrics, up from less than 20 percent in 2017.” In addition, we’ve seen that most of large growth in the monitoring space is occurring with new vendors, focused on business-relevant metrics (typically in the APM space).
Am I ready for the shift?
For IT leaders out there wondering if they are on top of this change, here are a few pointers to help make the transition.
Start with customer experience
Monitoring solutions must now start from application health and customer experience—how fast is it loading, how quickly could I complete my transaction—rather than beginning with the infrastructure layer.
APM is core
Application performance monitoring, log analytics, and UEM now become a core piece of your monitoring solution. They are closer to the pain, and closer to the user experience.
Don’t neglect the infrastructure
Having said that, ignore the underlying hardware or cloud services at your peril. Just because it’s not the top of the monitoring pyramid any more doesn’t mean it can be ignored. This is critical for problem resolution.
Combine when possible
So, what’s an IT leader to do? Tool sprawl has become a real problem and it’s only getting worse. Many of the visionary new vendors are providing great monitoring tools but can’t provide the single-view solution that the large vendors of the past could. It requires an additional effort to consolidate tools when possible and find ways to expand the aperture of new tools.
So, what can we learn from the good old days? We know that in the IT industry we have a fantastic cycle of: A. Solving a pressing problem effectively. B. Growing the complexity of that solution to meet a growing ecosystem. C. Radically simplifying a solution when the level of complexity becomes unmanageable.
The IT monitoring space is just getting to C.