Not too long ago, I read a piece in The Register that was cleverly titled, “AIOps they did it again, played with your heart, new acronym shame.” As you might imagine, the article sought to examine AIOps from all angles (non-believers, evangelists, and the unaware), but there was one paragraph, in particular, that really stood out:
“When you ask most people in the industry about the term AIOps, they roll their eyes. While it sounds cool, they collectively admit that the name over-promises. As Gohring told me: “Many vendors are confused by the term and end users are wary of it in the same way they’re wary of the more general AI/ML concepts – they want to understand how much of it is marketing BS and how much of it might be of value to them.”
I realize that I’m a marketer, but I assure you that AIOps isn’t “marketing BS.” In fact, if properly implemented, AIOps is designed to be transformational.
I understand if you’re still a bit skeptical, so here are four specific issues that AIOps resolves:
Problem #1: Identifying business service impact
Identifying the impact to a mission-critical business service as dynamic changes occur within related applications and their underlying infrastructure is really hard to do in ephemeral IT systems. Modern IT must continuously optimize workloads and processes while also mitigating the risk of unintended consequences to primary, secondary and tertiary systems.
The right AIOps plan allows you to take a business service view comprised of dependency and topology maps that stitch multiple applications and infrastructure layers — spanning multiple data centers and clouds — with real-time context, allowing you to visualize service impact and accelerate your mean time to repair (MTTR).
Problem #2: Accelerate your APM reach
APM reach can be defined in two ways. Reach that extends application visibility deep into the underlying infrastructure and reach that extends monitoring visibility to the vast number of applications currently left blind. It might not come as a surprise that traditional APM tools are so expensive that most organizations limit their deployment to on average 4-5% of their applications. As a result, organizations are suffering from gaps in their application visibility and are left guessing about the status of most of their applications’ health.
With an appropriate AIOps strategy, you’ll employ a mix of a fully instrumented (code level inspection) solution for custom applications with a light-weight application monitoring solution suited for the majority of off-the-shelf applications. In either case, the solution will use ML to apply topology and dependency maps to help build a complete view of the application health.
Problem #3: Automating incidents with an accurate CMDB (Fulfilling your ServiceNow ITSM vision)
According to recently released EMA research, 90% of early adopters stated their CMDBs are either extremely or very important to AIOps. And it’s no wonder, the CMDB needs accurate, real-time data to recognize what exists within an ephemeral environment. If the CMDB is not continually updated, or if the configuration item doesn’t exist in the CMDB, then it’s impossible to automate the incident process accurately, and you’re left trying to discern what’s actionable vs. what’s noise.
By improving the accuracy of the CMDB with a robust AIOps platform, you can not only automate the incident process but also automate other ITIL processes such as change and problem management that are built on the CMDB.
Problem #4: Taming Cloud Sprawl
IT organizations have struggled with containing cloud consumption and costs that emanate from orphaned workloads, unauthorized usage, or unsanctioned projects. Since multiple parallel organizations have direct access to cloud(s), they often spin resources up or down without ITs awareness, which often leads to increased costs. As a result, IT wants to identify unauthorized and underused resources to control their sprawl and mitigate the risk of wasted costs.
A true AIOps solution gives you multi-cloud visibility by accessing multiple data sources so you can control costs, sprawl, and enforce policies that alert you to what’s been requested, who requested it, when it was requested, where it was requested, whether or not the request was authorized, how it’s performing, and if it’s still in use.
Still not a true believer? If you plan on attending the Gartner Symposium/ITxpo 2018 in Orlando, stop by booth 535, and I’ll personally walk you through the promise and benefits of adopting the right AIOps solution for your organization.