From NoOps to AIOps, the IT industry is abuzz with buzzwords focused on automating IT operations processes. Work that used to involve lots of time and toil can, we're told, be automated away by algorithms and AI.
But that doesn't mean automation will dominate every part of IT operations work. There will always be tasks that even the smartest AI can't handle.
To prove the point, here's a look at six ITOps skills that only humans can wield effectively.
1. Policy Configuration
Policies — meaning files that define how resources should be configured or how processes should be completed — have become one of the cornerstones of modern IT automation. By writing infrastructure-as-code (IaC) templates and policy-as-code frameworks, IT teams can automate and scale complex configuration workflows that would otherwise require enormous amounts of manual effort.
However, someone has to write the policy files that drive these automations, and that somebody is likely to be an ITOps engineer. Algorithms and AI tools can't write policies very well because there are too many variables to consider and too many different outcomes and use cases to support or optimize for.
2. Incident Response
Automation has simplified the domain of incident response considerably over the past several years. Incident response platforms can now automatically coordinate response roles, drive communication between stakeholders, track response operations, and so on. Response processes can also be preplanned using playbooks.
That said, it's virtually impossible to automate incident response completely. Significant human effort will always be required to assess complex incidents and formulate a response plan. Humans also need to step in when playbooks go awry — as they tend to do, given the impossibility of anticipating all incident variables and conditions ahead of time.
3. Complex Scaling
Infrastructure and application scaling operations can often be automated. ITOps teams can write autoscaling policies that govern when additional nodes spin up within a managed Kubernetes cluster, for example, and they can rely on orchestration tools to scale up application instances automatically based on demand.
But sometimes preconfigured scaling policies and auto-management features don't suffice on their own. Applications may experience a sudden change in load that exceeds anything engineers had anticipated when they wrote autoscaling policies, for example, requiring a human to step in to provision additional infrastructure manually. Orchestration services can make mistakes, too, triggering conditions that humans have to fix manually.
4. Supplying Feedback for Application Enhancements
It would be great if there were a tool that could automatically determine which features or enhancements would help to optimize an application.
But there's not. The only way to decide how to improve an application is to collect feedback from people who know its flaws.
In this respect, ITOps engineers are uniquely positioned to help plan application enhancements because they're the ones who manage applications and troubleshoot problems on an everyday basis. They know which performance limitations should be priorities for developers to improve, or which new features might contribute most significantly to a decrease in end-user experience problems.
5. ITOps Tool Deployments
ITOps teams can leverage a variety of tools — like observability suites, orchestration services, and release automation software — to make their jobs simpler. These tools automate key parts of IT operations workflows.
But tools like these typically don't deploy themselves. Even if they're software-as-a-service (SaaS) solutions — as many are today — that IT teams don't actually need to install, the tools must still be configured to conform with whichever environment and resources they need to support. Although some vendors do a good job of auto-configuring their tools or offering good default configurations, some amount of manual configuration is almost always necessary. And that configuration work, of course, falls to IT operations teams.
6. Assisting End Users
End-user support — which includes tasks like assisting users who can't use the printer or figuring out why an application runs slowly on a particular employee's system — tends to be among the least glorious types of work that IT operations teams perform. But it's also among the most vital because no matter how fancy your IT systems are, they don't drive much value if your end users can't actually use them.
Because end-user support requests vary so widely in terms of what the problem is and what's required to fix it, they're almost impossible to automate. ITOps engineers need to be in the room, on the screen, or (at a minimum) remotely connected to the systems they are supporting, and you can't do that kind of work with a tool or algorithm. You need a human.
Conclusion
There's no denying that automation has transformed the way IT operations teams work over the past decade or two. Tasks that used to require tremendous effort can now be automated by tools such as IaC frameworks, orchestrators, autoscaling services, and more.
But at the end of the day, there remains a variety of ITOps tasks that are just not possible or feasible to automate. They require the skills of actual human beings, and they always will.
About the author
Christopher Tozzi is a technology analyst with subject matter expertise in cloud computing, application development, open source software, virtualization, containers and more. He also lectures at a major university in the Albany, New York, area. His book, “For Fun and Profit: A History of the Free and Open Source Software Revolution,” was published by MIT Press.
