Building an IT operations ecosystem that integrates from end-to-end provides many benefits, including easier automation, better data, and faster resolutions. Read this article to find out why you should employ end-to-end integration at your company.
If you started a new company today, would you prefer to buy different “best of breed” products for your social media, marketing, sales, and customer support needs and integrate them all after the fact, or would you invest in a single solution that pre-integrates multiple services with end-to-end functionality?
The answer is simple: A pre-integrated solution provides advantages in cost and (perhaps more importantly) time to value. In addition, there are benefits in the smooth processes and reduced implementation work that such an out-of-the-box solution provides. However, the real value is in actionable data.
Clean, aligned data is much harder to get when working with multiple cloistered systems. With clean data, you get better analytics, which drives continuous improvement. Unfortunately, most companies don’t get a “start over” after attempting a best-of-breed approach, and the inertia that comes with implementing legacy systems is difficult to counteract.
No more apparent is this challenge than IT management
When we explain our vision of IT management with IPsoft clients and partners, we focus on the benefits of end-to-end execution of tasks without human involvement, i.e. intelligent automation or autonomics. Our Amelia AIOps (formerly 1Desk) platform serves as a catalyst to drive a conversation with IT Operations teams around continuous improvement fueled by data, analytics and process.
- For Further Reading: Autonomics Enhances Your Previous IT Investments
The ultimate goal of AIOps is to fix everything in IT environments automatically and if human engineers are required to be involved, the system monitors, learns and extracts data from their manual efforts. AIOps reimagines how people, process and technology can be re-architected to achieve a fully automated IT operation – and it is only possible because the platform is built with an end-to-end perspective in mind.
The purpose of AIOps is to generate actionable data for continuous improvement
As part of our discussions with customers and partners, it is critical to define what we mean by “end-to-end,” i.e. where does it begin and where does it end. Let’s take a standard IT Operations use case.
Some people may believe that it starts with monitoring of systems, applications, etc. However, it actually starts before, as we need to have information about the system, which is usually found in the CMDB, i.e. discoverable info such as configuration, relationships, dependencies as well as business-related information such as owners, vendors, standard operating procedures, etc.
AIOps reimagines how people, process and technology can be re-architected to achieve the vision of a fully automated IT operations
There are lots of different types of monitoring, e.g. logs, performance, configuration changes, etc. Different monitoring systems will generally send events into a central place where intelligent correlation engines and anomaly detection algorithms are constantly analyzing and determining whether an issue or problem should be investigated, and if so a ticket is created in the ITSM system.
Once a ticket is created, automation should kick in to further diagnose the situation, determine root-cause, and if a change is required, open a change request so that an automated remediation can be executed. This is usually where an ideal support story ends, otherwise appropriate human engineers need to get involved.
If that’s the case, assigning the ticket and alerting the engineer can be automated or done manually with various degrees of intelligent routing. An engineer would have access to diagnostics performed by the system and would eventually resolve the issue and close the ticket. And this is where the majority of support stories would end.
However, this is not where AIOps ends — we record the steps human engineers take to resolve issues and automatically build new automations based on those steps.
In order to automate a fix without AIOps, most companies today would have to tie several systems together, from CMDBs to multiple monitoring tools, to event correlation system, to ticketing systems, to notification systems and finally to automation solutions. Some of these systems require two-way communication. These integrations aren’t easy to build, and although they’re certainly possible, they can present future challenges. What happens when you need to upgrade one of the applications? What happens when one of the integrations stops working? At that point, every step of your automation is temporarily null and void because one element of the integration has stopped working. Also, when integrating disparate products some items can be "lost in translation."
AIOps is designed as a learning system by observing what human engineers are doing to remediate a problem. So every single time a ticket is closed manually, AIOps generates a new ticket with as much information as possible, so it can be analyzed and correlated with previously resolved issues. Now the team responsible for permanently fixing the issues have actionable information to build or rebuild intelligent automations, and if there are questions, the team can always contact the operations engineer that initially resolved the issue and get additional insight if needed. The goal is that every manually performed resolution is recorded, analyzed, and utilized to build new automations.
The key here is a closed loop for a continuous improvement cycle, using all the data collected for analytics and insight. Many vendors talk about “end-to-end,” but define it differently — and that matters when it comes to overall efficiency and effectiveness. For us, the final “end” for AIOps is automating with Digital Labor – and creating a collaborative environment of where virtual and human engineers can excel at operational efficiencies.