Observability, Availability, Scalability and White Box Analytics

Understanding the latest marketing nomenclature in operational analytics

The word ‘observability’ has been thrown around quite a bit in measurement circles over the last couple of years. In this blog, I will describe what observability means, how it relates to availability, how it impacts scalability, and how IntelliMagic Vision fits into the observability landscape.

What is Observability?

Observability is a measure of how well internal states of a system or application can be inferred from knowledge of its external outputs (data, signals). This stands in contrast to monitoring, which is something we actually do. Observability (as a noun), is more a property of a system, application or service.

For the sake of this blog I will refer to a system, application or service as a thingy.

You can ask if a thingy is observable, but you can’t observable a thingy.

Being Known

When I discuss observability with friends, I think it is helpful to use the analogy of human interaction. Humans can relate to each other at various levels just like thingys. Humans use verbal and physical expressions to signal the outside world to what is going on internally in their body, mind and soul.

Thingys either push their inner state or allow the inner state to be known via external polling (i.e. – monitoring). Whereas thingys can be designed to be observable, humans are all designed to be observable but may either choose not to be observable or may lack the ability to be observable due to physical, mental or emotional limitations. In human terms, we can think of observability as the ability to be known.

People who study human behavior or who have special gifts of intuition may be able to consume limited signals such as minor facial muscle movements to decipher some of person’s inner working. This can provide some insight into the inner person but for something as complicated as a human, one would require background, shaping experiences, outlook and beliefs just to scratch the surface of understanding someone.

While thingys are not as complicated as humans, thingys and/or the architectures that bind them together continue to become more complex and abstract.

IT Availability: The Ultimate Goal

The ultimate goal of observability is to maintain IT availability. Understanding the inner workings of thingys allows us to maintain a healthy operational state so that IT services can be delivered without interruption.

Just like humans that only function within certain constraints, thingys also have basic requirements. Whereas humans need certain levels of caloric consumption, temperature ranges, oxygen levels and healthy human interaction, thingys can only function well when they have sufficient compute, memory, storage and connectivity to other thingys they are functionally related to.

At a minimum, these external indications of the health of both humans and thingys need to be sufficient in order to understand why a functional failure occurred. In IT we call this root cause analysis.

Just as in the study of both humans and thingys, root cause analysis is not nearly as helpful as proactive or predictive health optimization. Ideally the same indicators of health that were present while the thingy was operational can be used to proactively correct the thingy before it becomes unavailable.

The Scalability Challenge

Improved observability will provide access to an increasing amount of data. This data will need to be consumed intelligently in order to both scale and leverage the ever-increasing number of data points.

If you have been in the monitoring or measurement world for any amount of time, you have hit scale issues. Scale issues directly relate to the growing number of data signals from the thingys that we are trying to understand. How do we deal with the scalability issue?

White Box Analytics

One of the techniques that can be used to address the scalability issues resulting from observability is white box analytics.

In the DevOps space, white box analytics has referred to analysis of signals produced by applications. The DevOps definition does not address the scale issue resulting from observability. At IntelliMagic, we use white box analytics to analyze thingys.

White box analytics use deep understanding of the type of thingy to synthesize millions of signals and provide deep understanding of a specific instance of a thingy. At IntelliMagic, best practices and expertise are embedded into the signal reduction and transformation processes resulting in refined information that can be used to proactively inform on the health of IT thingys and improve the availability.

IntelliMagic = Observability + White Box Analytics = Availability Intelligence

If your enterprise IT teams are struggling with the challenges relating to observability, please let us help. We provide solutions to improve the availability of the complete enterprise mainframe stack as well as the distributed storage I/O stack.

For a list of supported mainframe thingys please refer to: intellimagic.com/mainframe

For a list of supported distributed thingys please refer to:  intellimagic.com/san