Observation has recently become a strong word and is making headlines in DevOps and IT publications. Industry experts, such as charity leaders, Honeycomb’s technical director and co-founder Cindy Sridharan and others, spread the message about the importance of visibility and make it clear that this is more than just a passing trend, it is an approach that development organizations should follow.
The increasing complexity of software systems is one of the main reasons why observability is much more than a buzzword or a trend. Daniel Spoon Spoonhower, co-founder and CTO of Lightstep, recently described the growing need for visibility, noting that whether organizations move to micro-services based on Kubernetes or use AWS Lambda, all teams rely on software they have not written or controlled. We probably all use hosted storage systems to build cloud projects, or use open source systems managed as part of the production and implementation process, or rely on third party APIs to build applications. Spoonhauer explained that the increasing integration of these new processes and platforms leads to what he called layers of different properties, requiring teams to understand the relationship between all these interdependent components.
What is observation?
The term perceptibility originated in control theory, perceptibility refers to the way in which the internal states of a system can be derived from the external outputs. As our software systems are distributed across an increasing number of platforms, tools and teams, development organizations are building visibility into their systems to achieve both comprehensive and granular visibility.
James Governor recently stated in a RedMonk article that monitoring crews helps to understand both the overall condition of the system and the condition of all components, making troubleshooting easier. As mentioned in HoneyComb’s Guide to Achieving Observability : Observation is the ability to ask random questions about your surroundings without having to – and it’s the key to knowing in advance what you’re going to ask.
Observation versus monitoring: Which perception is not.
Another way to understand perceptibility is to distinguish it from monitoring. Many tend to mistakenly believe that surveillance is a less effective precursor to perceptibility, perhaps because the term surveillance was popular before perceptibility became a buzzword, but this is not the case.
Cindy Sridharan explains that monitoring and control are two different and complementary practices. Although monitoring provides a panoramic view of the performance and behaviour of systems in nature, it can help us understand the weaknesses and changing needs of the system and, according to Sridharan, is best suited to report on the overall state of the systems. Observation, on the other hand, aims to provide a very grandiose insight into the behaviour of the system and a rich context that is ideal for problem solving.
Developers are doing their best to reach a wider audience Download the free report
The three pillars of perceptibility
Many define perceptibility as the sum of three key elements: Metrics, newspapers and distributed spores:
# Metrology is a numerical value measured over a period of time. They can help us determine the status of our system and are often used to set off alarms.
# Minutes are temporary reports of events that can give us a detailed account of what happened in our application.
# Distributed dependencies are representations of the execution of the code. They help us see the end-to-end execution process through a distributed system.
Although these three tools help us to better understand our systems and to find the cause of the problem, some argue that this understanding is more than the sum of these three parts.
Brian Cox notes that the purpose of the observer team is not to collect logs, measurements or tracks. The aim is to build an engineering culture based on facts and feedback and then spread this culture throughout the organisation. Mr. Sridharan added that observation is not about logs, measurements or traces, but that data is managed during troubleshooting and feedback is used to iterate and improve the product.
Ben Siegelman, CEO and co-founder of Lightstep, also disagrees with the idea that perceptibility is simply a combination of measurements, magazines and distributed footprints, arguing that none of them responds directly to a specific hotspot, use case or business need. When it comes to perceptibility, it’s what you do with the data that counts, not the data itself, says Siegelman.
Perceptibility: A look at our distributed systems
The more complex systems become, the more difficult it becomes to get a clear picture of what is going on or to get a complete picture of the behaviour of the system. Our systems are constantly evolving and it is no longer enough simply to browse through dashboards and sort logs to solve problems as soon as they occur.
Spoonhower lists describe measures of success for organizations that utilize perceptibility, improved and even more frequent implementations, faster call response by reducing the time it takes to understand the root cause of a problem and resolve it or even solve problems before they occur, and finally, optimized performance to deliver a better user experience for customers.
Reliable and frequent use, fast troubleshooting and a positive user experience are the successes we all strive for. Adopting the right tools, processes, mindsets and teamwork will help organizations to take the DevOps game to the next level.
*** This is a syndicated network of security bloggers from Blog – WhiteSource, written by Ayala Goldstein. You can read the original announcement at https://resources.whitesourcesoftware.com/blog-whitesource/observability.observability definition,observability instrumentation