The debate continues about whether NoOps — the idea that IT processes and infrastructure can be so completely automated that the IT operations team becomes unnecessary — is actually an achievable reality.
But even if you can't totally eliminate IT operations, you can significantly reduce the amount of toil and tedium ITOps teams have to exert to manage key infrastructure and workloads.
Perhaps nowhere is this more true than in the cloud, where a litany of modern automation tools makes it easy for IT teams to come closer than ever before to making NoOps a reality.
Read on for a look at different types of cloud automation tools that help enable a NoOps strategy, along with an assessment of how close they come to making NoOps work in practice.
Tools for Achieving a NoOps Strategy in the Cloud
Cloud automation tools that lend themselves to a NoOps strategy can be broken into different categories. Let's explore them one-by-one.
Cloud recommendation tools
Each of the major public cloud providers now offers automated recommendation services designed to help businesses optimize their cloud workloads from cost, performance, security, and other perspectives. AWS offers Trusted Advisor, Microsoft Azure has Azure Advisor, and Google Cloud Platform (GCP) offers Google Cloud Recommender.
These services automatically scan cloud environments and make suggestions about how to improve workloads. They can also apply the recommended changes automatically in most cases.
Cloud recommendation tools don't totally achieve NoOps because someone has to review the recommendations and click the Apply button. They're also not capable of detecting every possible deficiency with a cloud workload, so manual auditing will always be necessary. However, these tools can significantly reduce the amount of time and effort required to align cloud workloads with best practices, and in that sense, they bring IT teams one step closer to implementing a NoOps strategy.
NoOps and cloud autoscalers
Another type of cloud automation tool that can dramatically reduce IT operations work are autoscalers. Businesses that use cloud services like virtual machines and managed containers can set up autoscalers to increase or decrease the size of their cloud infrastructure automatically. In some cases, the autoscalers leverage machine learning to scale as efficiently and intelligently as possible. (Without machine learning, autoscaling is usually based on preconfigured policies and thresholds that define when scaling should happen.)
The alternative to autoscaling is to monitor and adjust cloud infrastructure manually — which is exactly the opposite of what you'd be doing if you were pursuing a NoOps strategy.
Code generation as a step toward NoOps
Generative AI engines that can automatically write code, such as GitHub Copilot and Amazon CodeWhisperer, have garnered attention primarily for their potential to help automate software development.
But these tools could enable a NoOps strategy, too, by automatically writing the code that IT operations teams would otherwise need to produce by hand to perform various maintenance and management tasks. For example, instead of creating infrastructure-as-code templates by hand, ITOps teams could use a tool like Copilot to write the code for them.
Humans would need to initiate and manage the process — and they'd probably need to tweak the code, given the current limitations of code generation tools — but they'd still be doing a lot less work than they would if they wrote everything from scratch and by hand.
Automated cloud failover
When something goes wrong in the cloud, IT operations teams normally have to respond in a highly manual way. Given the unpredictable nature of outages, it's hard to automate recovery operations and incident response.
But cloud providers are working to make incident management easier to automate. For example, Google Cloud says it can perform automated health checks for workloads and reroute traffic to alternative endpoints in the event that one workload instance fails.
Solutions like this wouldn't make cloud operations a totally hands-off, NoOps affair. Even with automated failovers in place, someone would need to figure out what caused the initial failure and how to prevent it from recurring. But this type of solution helps to reduce the urgency of real-time incident response, since automated failover helps to ensure that services remain available in the event of a partial outage, without requiring human intervention.
Conclusion: The March Toward Real-World NoOps Strategies
None of the tools or techniques described above enable true NoOps — perhaps because the idea of completely automating IT operations just isn't realistic. But they do show that placing a NoOps strategy at the heart of cloud management is more practical than you might think. The state of cloud automation now makes it possible to operate much more efficiently, and with much less toil, than many people could even imagine when the NoOps conversation began in the early 2010s.
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.
