• John Atkinson, Director Solutions Engineering, UK & Ireland at Riverbed Technology


• 16.03.2023 09:00 am
#cloud #data

The role of observability is changing. Traditionally, it has been viewed as a way to help DevOps teams tackle the challenges they experience in complex, highly distributed, cloud-native environments. In this context, observability has evolved from application performance monitoring to solving broader challenges across application monitoring, testing, and management in these environments. Consequently, there needs to be a wider definition that applies to this wider set of challenges: Unified Observability.

It can be difficult to execute Unified Observability, especially for global enterprise organisations. Consider, for example, a company with 10,000 members of staff and all these employees will anticipate dependable and robust digital experiences. The challenge is, they will often be operating in rapidly evolving hybrid working environments, each with different WiFi and laptop configurations, and they will expect just as strong a digital experience in each as they would have on-premises. To complicate matters further, this all comes before accounting for the potentially hundreds of thousands of customers using an unknown mix of cloud applications, legacy on-premises, and shadow IT. In such instances, a Unified Observability solution that can collect full-fidelity data from all sources while overcoming siloes and locales is crucial.

Leverage fewer tools for more effective outcomes

90% of all IT teams are harnessing observability tools to achieve visibility and successfully manage a variety of geographies, applications, and networking needs, according to recent research commissioned by Riverbed and undertaken by IDC. What’s more, approximately half of those groups leverage six individual observability tools, which produce an unmanageable number of daily alerts. The sheer number of alarms is compounded by the high volume of data these tools generate, meaning it’s it incredibly hard to establish if all the important information is gathered. 

The issue is exacerbated by the use of outdated or limited tools. Of IDC’s respondents, nearly two-thirds reported that the tools used by their organisations focused exclusively on the business' complex layers of hardware configurations, cloud-based services, and legacy on-premises applications. And 61% of IT teams said they believe this restricted view stood in the way of both collaboration and productivity.

It is here that Unified Observability tools come into their own, by empowering IT teams to amalgamate full-fidelity telemetry from across devices and domains, as opposed to collecting and sampling snapshots of data which can result in considerable gaps. In the same way that if an organisation only collected a small portion of customer complaints, they would not know the full extent of issues, making it impossible for them to resolve them, and therefore causing a significant loss in customers. In both cases, the key problem is that samples can’t give a full picture which can result in an incomplete picture and lead to the wrong conclusions. 

Use automation to minimise noise

To enable IT teams to make sense of the data to provide actionable insights, they are starting to combine Unified Observability with artificial intelligence (AI) and machine learning (ML) to detect any anomalies much more quickly.

A key area this is being leveraged for is alert management. Manually sifting through a large volume of alerts, to identify the root cause of problems, is highly time-consuming and can leave IT teams feeling worn out. Especially when they are also trying to keep up with the constant flow of data from full telemetry. Historically, resource-intensive war rooms were turned to for this process but they tended to be inefficient and rarely solutions orientated. The alternative approach of having a senior-level employee identify individual problems across IT siloes wasn’t much better. It was a waste of resources and resulted in over-dependence on a singular or small group of people, which quickly became problematic if they decided to leave the company with no way of producing the same results.

However, with Unified Observability, the number of tickets and alerts that come to IT teams is cut to a minimum and siloes between teams are overcome, facilitating cross-team collaboration. This drives up efficiency and simultaneously job satisfaction. In light of ongoing talent shortages, Unified Observability is, therefore, a vital resource for reducing the daily pressure on IT teams.

Streamline processes with Unified Observability

In addition, it has been a standard for IT teams to have documented runbooks for the manual resolution of specific problems, but ML and AI can be harnessed for the creation of runbooks that automate these processes and simplify the identification of solutions. For example, through the creation of customised, preconfigured solutions for frequently encountered issues. In addition, these engines can be personalised, enabling teams to edit them until they are certain they will result in positive outcomes. This frees up the time of expert staff, allowing them to focus on strategic tasks.

Enhancing productivity

The recent IDC survey sponsored by Riverbed uncovered that 75% of teams find it challenging to glean insights from the siloed variety of observability tools they use. These pain points can be avoided with Unified Observability which empowers IT teams to analyse the complete span of their organisation’s data to develop actionable insights. These insights can subsequently be used to guarantee a seamless digital experience for end-users whilst ensuring employees remain highly productive. All while automated remediation increases agility, optimizes costs and improves services, at the same time.

There is a widespread understanding that monitoring infrastructure is crucial to deliver seamless digital experiences for employees and customers, as illustrated by the rising number of organisations embracing observability solutions.  However, numerous companies are continuing to use a plethora of outdated legacy tools which have limited capabilities and do not provide a comprehensive view of network performance and end-user satisfaction. More and more organizations are now turning to Unified Observability, the optimum solution to reduce toolsets and glean actionable insights to improve organisational productivity and accelerate business performance.