In 2020, DevOps and AIOps go hand-in-hand

Today’s DevOps professionals have a lot to monitor and react to in order to keep IT systems running including high volumes of alerts, signals coming from disparate tools and considerable amounts of IT noise. What’s more, their workload has only increased since the shift to remote working due to COVID-19. By Guy Fighel is GVP & Product GM, Applied Intelligence at New Relic.

  • 3 years ago Posted in

On top of all that, they are expected to continuously improve IT infrastructure performance, problem-solve more accurately and find incident resolutions more readily. So much data and so many alerts can cause response fatigue and make it hard to prioritise issues and know what they really need to act on.

Busy DevOps engineers and leaders are well aware of how AI can help them achieve their goals, possessing a keen ‘automation mentality’, whereby they identify opportunities to automate away toil  by deploying a tool and thus save time down the line. They recognise that the more time they save with AI taking on manual tasks, the more time they have to spend focusing on more complex and higher-value tasks.

Research by New Relic and Vanson Bourne revealed that out of 750 global IT decision makers, 89 percent said they believe AI and machine learning (ML) is important for how organisations run IT operations, and 84 percent also remarked that AI and ML will make their job easier. Plus, findings from Gartner shows use of tools such as AIOps specifically is growing – it predicted their integration in large enterprises will grow from 5 percent in 2018 to 30 percent by 2023.

AIOps tools that detect, diagnose and resolve problems and improve incident response are vital to the success of today’s DevOps professionals, but in what ways is AIOps helping them exactly?

1.     Automatic anomaly detection

Some of the latest AIOps tools automatically monitor and detect anomalies via site reliability engineering golden signals such as latency, saturation and traffic. They can then send notifications to IT teams including details about the anomaly. This enables them to quickly and easily assess how to respond, before it potentially causes a problem.

2.     Data-agnostic tools for richer data analytics

Data-agnostic AIOps tools allow DevOps teams to leverage data from numerous sources; standardise it and improve its usefulness with metadata to provide greater context, such as which components are related. This allows users to have a greater understanding of the problem and thus reach the root cause of any issue faster.

3.     Correlation of related incidents to reduce IT noise

DevOps teams are used to noisy environments, but AIOps helps them significantly reduce large volumes of alerts down to manageable amounts and thus avoid alert fatigue. This is possible due to AI establishing relationships between cases of incidents that are alike or related. Some tools also become ‘smarter’ the more they are used, enabling the user to feedback to the AI, for example, by confirming that it correctly identified alerts were resulting from one issue, training it to spot similar instances in future.

4.     Augmentation of incident management

The use of AI is not to replace those working in DevOps, it’s to augment routine activities so that workers can perform better. The two working in tandem together means organisations get the best of both the ability to manage huge datasets accurately from AI, enhanced intuition, and the  combined decades of experience of the people that make up the IT team doing their jobs. There are AIOps technologies that include ‘decisions builders’, which allow users to create their own logic based on event attributes or choose similarity algorithms out-of-the-box to correlate incidents. Tools that are transparent rather than opaque also allow humans to stay fully in the loop with why certain actions were taken so they can stay in control of the process and avoid missing critical signals.

5.     Accurate routing of incident for ownership and actioning

AIOps tools can automatically suggest where to route incidents based on data about the issue and enable DevOps professionals to improve the process by which tasks are distributed among the team. For example, they can mark cases related to a specific application to be sent to a dedicated group, and if they already have too much on, go to another team member with relevant experience and the capacity to own it.

Those in DevOps today may be experiencing the busiest work lives they and their colleagues have ever faced in their careers right now, particularly since the shift to remote working. At the same time though, they have the most advanced technologies at their disposal to deal with the high volumes of alerts and disparate signals successfully. AIOps tools truly go hand-in-hand with DevOps. This means IT professionals possess quicker and easier ways to identify issues, create diagnoses and find the right resolutions to issues, both after and before they cause problems.

 

By Steve Young, UK SVP and MD, Dell Technologies.
By Richard Chart, Chief Scientist and Co-Founder, ScienceLogic.
By Óscar Mazón, Senior Product Manager Process Automation at Ricoh Europe.
By Chris Coward, Director of Project Management, BCS.
By Trevor Schulze, Chief Information Officer at Alteryx.