See, Enforce, Automate: A Structured Approach to Observability

On trend in market right now is the importance of going back to basics. Whether this is due to the global pandemic, other international events of late or just a collective feeling that everything has become a little bit too complicated, there's a clear desire to focus on first things first.

For the most part, people want simplicity in there lives. Want proof? Look no further than the (unbelievably true) statistics found in this 2022 Nordpass Research report. It says among the top 50 most common passwords used by C-level executives, managers and business owners, "123456" is raked #1, and it's the reason for over 1.1 million cybersecurity breaches worldwide. While we crave simplicity, many times the basic IT necessities, like password integrity, can be overlooked.

When it comes to automation (and orchestration) we must take this lesson to heart. Keeping it simple is the key to realizing business value, like increased productivity, reduced cost and greater security. An automation solution that is complex or without a clear level of visibility puts you on the fast track to chaos.

First step in a strong observability strategy: cover the basics

At WWT, we abide by the philosophy, "less haste, more speed," which is why, when it comes to automation, we look to first address the basics of observability. You cannot protect what you cannot see. So, through our assessment process, we work closely with your team to uncover maturity levels and identify any gaps, starting with operational readiness. This includes elements such as:

• CMDB assessment
• Tools rationalization
• Infrastructure dependency mapping
• Application dependency mapping
• Public cloud monitoring

It is at this preliminary stage where so many businesses can slip up. There is a whole host of reasons this happens, not least because it is expected that the basics are in place – in just the same way it's assumed that people aren't really using 123456 as a password!  

Data democratization is simpler than it sounds

Many times, IT programs are underfunded or incomplete due to a limited understanding of breadth and importance of discovery and data accuracy. System complexity and a lack of clear data also makes it difficult for business stakeholders to utilize the necessary information to do their jobs and make key business decisions. This can lead to an over-reliance on already stretched IT personnel to spin cycles interpreting native data.

In achieving an effective observability strategy, an organization should first aspire to establish a state of data democratization, enabling everyone to understand and process data no matter their role or position. This puts the most relevant stakeholders back in control and empowers an organization to act with agility.

Here's a good example of why data democratization is a useful strategy: Your end users need, and rightfully expect, a seamless technology experience when performing required daily functions. Let's say they make complaints about poor network connectivity. Without an effective observability solution in place much time and effort will be exhausted sifting through data to find the root cause analysis. Weeks later it's determined that the network is fine and it was a low-latency application issue that caused the issue. See where I am going with this?

The right observability strategy would ensure clear context of data flows in advance of an incident. Aggregation and correlation of the multitude of telemetry sources in place across cloud services, applications, and middleware will help you realize many benefits, including effective event correlation, improved route cause isolation and lowered MTTR.

In addition, data democratization puts the organization in a position of maturity concerning monitoring systems and event triage, making it easier to integrate more advanced solutions for incident prevention, machine learning/AI, automation and orchestration.

The 5 levels of maturity

There are various dependencies and workstreams within WWT's foundations assessment. It is fairly typical for us to find an organization to have relatively advanced areas, while others have been neglected. Here is how WWT frames the levels of observability maturity within an organization:

Level 1 – Chaos

The organization has an ad-hoc approach to managing data flows with limited or non-existent tooling. Users detect IT issues before IT who are further impeded by multiple helpdesks and a lack of documentation.

Level 2 – Reactive

Core technology components are monitored through the use of multiple, disparate tools. Incident triage and correlation happens manually.

Level 3 – Proactive

Standardized tools are in place with effective performance monitoring. Predictive analytics are used and proactive detection means IT are usually aware of a problem before the end user.

Level 4 - Service 

Tools and processes are service-aligned and operations has a single pain of glass approach. There is context from business to IT Ops, and data democratization empowers business stakeholders. Real time corrections and self-healing lead to reduced business impact.

Level 5 – Value

An organization can drive true automation and orchestration via a coherent set of data, process and tools. Individuals are presented with actionable intelligence with insight into business not just infrastructure.  

Simplify the complex

Unless the primary building blocks are addressed, it becomes almost impossible to realize the value of more complex themes. Therefore, deploying a structured assessment process not only makes it easier to identify those gaps but also enables an easier means of explanation to board-level sponsors.

Taking a look at the diagram above, you see the many different data source types that need bringing together if an AIOps approach is to be successful. Applying observability early on as a basic principle will help you see, understand, and optimize the movement occurring inside and beyond the IT architecture, all with the benefit of supporting the business side of your organization.