There is a tendency to confuse MLOps and AIOps. While there are some common characteristics between the two, MLOps and AIOps are two different domains, are applied differently, and serve different goals.
Let's start with the definitions.
AIOps, sometimes referred to as Artificial Intelligence for IT Operations or Algorithmic IT Operations is a term invented by Gartner in 2016 as an industry category for Machine Learning analytics technology that enhances IT operations analytics.
Gartner's official definition of AIOps is the following:
AIOps platforms utilize big data, modern machine learning and other advanced analytics technologies to directly and indirectly enhance IT operations (monitoring, automation and service desk) functions with proactive, personal and dynamic insight. AIOps platforms enable the concurrent use of multiple data sources, data collection methods, analytical (real-time and deep) technologies, and presentation technologies.
In other words, Machine Learning and Big Data are the pillars of AIOps.
AIOps' ultimate goal is enhancing IT operations. The central process in AIOps is the intelligent filtering of signals our of the noise in IT systems. We can understand the problem that AIOps solves as follow:
Logs, traces, tickets, incident data, system configuration status, and any information related to a given system's performance and operation are not intended to be the goal in itself, but rather the tools.
When the amount of collected data that can be examined and used to draw conclusions exceeds a certain limit, our understanding, assumptions, and priority patterns become distorted. Most importantly, identifying root causes and suggesting solutions become slower and less accurate.
AIOps is an approach to use machine intelligence to solve this problem by collecting, processing, and observing intelligently IT operations data in order to identify root causes and propose solutions fast. In some cases, AIOps have the capability to solve problems without human intervention.
If you are familiar with DevOps, you will not find MLOps hard to understand. DevOps is a practice for collaboration and communication between developers and operations professionals to help manage production, decrease the time-to-market, and implement a culture of continuous testing, development, and feedback. MLOps has almost the same definition, except that developers here are usually data scientists, AI specialists, and Machine Learning engineers.
In other words, MLOps follows a path similar to that of DevOps. While DevOps focuses on shortening the product life cycle by creating better products each time, MLOps drives insights that can be put into better use immediately.
Compared to DevOps, MLOps is more experimental in nature. This new framework requires data scientists to try a variety of features, parameters, and models.
It is obvious from the two definitions above that the two domains are different and certainly not to be confused. Even if both share this willingness to make our systems better and more efficient, the two fields overlap but do not meet under the same umbrella, neither in terms of the approach nor in terms of the raison d'être.
While MLOps' goal is bridging the gap between data scientists and operation teams, therefore between ML model building and their execution, AIOps focuses on automating incidents management and intelligent root cause analysis.