Skip to Content

Internal system monitoring: The key to hosting success

Capgemini
2019-01-22

When you investigate world-class hosting providers, it’s easy to be taken in by pretty slides and presentations. You see the wonderful hosting experience you can expect, but more important details are frequently omitted.

At the end of the day, securing a physically safe location for your website is only part of the equation. If the majority of your site support work still falls on your internal staff, how valuable are the services in reality? Many organizations are unaware and shocked to realize this little detail after their new hosting contract has been signed.

In our view, what makes a hosting provider a valuable partner comes down to whether they provide system monitoring. System monitoring is the “secret sauce” that can make a difference between a successful or a failed hosting engagement. In many hosting situations, monitoring is either not provided or ends up being an expensive add-on cost. As a company, we feel quite strongly about this: not only do we provide internal system monitoring as part of our standard hosting engagements, but we also deliver data analysis and client access to our monitoring instrumentation at no additional cost. Transparency and diligence are what makes for a valuable hosting partner.

The Mechanics of Internal System Monitoring

Internal system monitoring can be best explained within the context of a hospital. Bear with me – it makes sense, I swear.

During a hospital visit, you’ve undoubtedly seen a nurse’s station. If you every wonder how a small team of nurses can safely handle a large number of patients you’ll quickly realize the nurse’s station is the key. Each patient ward has one, and with top-of-the-line monitoring technology, nurses can organize, prioritize, and optimize how they serve their patients. The system monitoring provided by LYONSCG is similar to the nurse’s station: we have a set of tools that allow us to bring all monitored data to a nexus for rapid alerts, analysis, or digital triage.

In each room, nurses use tools that acquire and transfer patient biometrics back to the nurse’s station. In our world, these tools are called monitoring agents and their purpose is to grab data from various sensors and pass it along to the monitoring server (aka, the nurse’s station). Our monitoring server acquires the data from these agents and then stores it for analysis and to guide any future action.

Our Toolkit

In a hospital, there are many types of diagnostic tools that provide visibility into general patient health. These include real-time monitoring of blood pressure, O2 saturation/respiration, heart rate, etc. Any anomaly seen with the instrumentation will generate an alert, and the nurses will immediately swing into action. In the system world, the parallels are striking and provide answers to similar questions like is the system running (O2 level), is the processor being overtaxed (blood pressure), and is the system available/alive on the network (heart rate).

These fundamental measurements allow us to determine the functioning health of our systems (or our patients). Knowing a system is up and running is great, but its only part of the digital health equation. These initial measurements tell us the patient is alive but doesn’t explain if they are safe.

Hosting commerce applications requires us to look deep into the inner workings of site health. In the medical world, this visibility can be gained using a tool like an electroencephalograph (EEG). Our monitoring tools are like the EEG, (noninvasive in nature), but instead of brain activity are designed to pick up operational fluctuations in site performance and indicate when there might be a problem. Data metrics based on application health, number of running processes, database replication state, processor load, network utilization, or memory consumption represent the current operational state of the environment, which is then mapped into a real-time graphical presentation and, when necessary, is used to generate real-time alerts.

Good monitoring tools are based more on a modified EEG technique called “evoked potentials,” and involves averaging the monitored activity into a time-locked presentation. In the world of internal system monitoring, we store our data in a time-series database, which allows us to use statistical methods to analyze and extract meaningful insights. These data points are broken down to identify current issues or to help predict future performance.

Internal system monitoring is one of the most important and valuable aspects of a successful hosting engagement. To stay alive and healthy, commerce applications need both contemporary site reliability engineering as well as a solid DevOps focus to enable service providers to evolve and optimize their hosting environments. This monitoring provides clients with the data needed to help make informed decisions and, most importantly, realize their commerce potential.

In conclusion, when you are looking to partner with a hosting provider, always ask yourself the following questions:

  • Do they provide internal system monitoring?
  • If so, what type of instrumentation and methodologies do they use?
  • What are they charging, and what’s all included?

Answers to these questions will help you make the most informed hosting decision possible and will help optimize your digital commerce performance.