DevOps / Machine Learning / Sponsored / Contributed

How to Turn Your AIOps Vision into Autonomous Cloud Reality

11 Jun 2021 9:00am, by

Saif Gunja
Saif is director of product marketing for Dynatrace’s cloud automation and DevOps solutions, bringing over 10 years of IT and marketing experience from his previous roles at VMware, Apple and Deloitte.

In most global organizations, DevOps teams have their backs against the wall. Even before the pandemic, business demands were soaring for new customer experiences, enhanced employee productivity and faster time to market. Now those calls for accelerated digital delivery drown out virtually everything else as organizations compete for post-pandemic differentiation and growth.

Yet DevOps teams are still restricted in their ability to support these requirements. Labor-intensive manual tasks, siloed tools and the complexity of dynamic, microservices-based environments, all complicate efforts to drive greater digital agility. However, as we at Dynatrace have seen from our own experience, there is light at the end of the tunnel for organizations that choose a path from manual operations to an autonomous cloud.

Moving Beyond the Daily Grind

Manual, routine tasks have always dominated IT’s attention, from provisioning infrastructure to conducting software integration testing. With limited practitioners on the ground and growing requirements from the business, the challenge has escalated over recent years — and the pandemic only exacerbated this dynamic. Particularly in the early days of the crisis, IT was forced to drop everything to support the rapid shift to remote working. One recent report claims that this “keeping the lights on” work occupied most (68%) of IT’s time in 2020, and only 37% of global IT teams were able to complete all the projects asked of them, versus 41% the previous year.

This work is time-consuming and repetitive. It spans from triaging operational problems to provisioning cloud servers. Even more time and agility are lost as DevOps teams struggle to gain insight into their increasingly complex application and infrastructure environments. Our own research has found that despite using an average of 10 different monitoring solutions, teams have observability into just 11% of their environments. Not only does all this work reduce the time that could otherwise be spent delivering new value to the business, but it also compounds silos between development, operations and business teams, which creates a further roadblock on digital innovation.

The answer is greater automation. According to our poll of CIOs last October, 70% said their teams spend too much time on manual tasks that could be automated, and that automation would reduce this wasted time by 38% on average.

Why AIOps?

The automation-first approach is often described as AIOps. The ability to deploy, monitor and manage all of your applications and the infrastructure on which they run automatically is particularly well-suited to the new reality of mass remote working. It’s also about making faster release cycles less error-prone and customer experiences fault-free through enhanced observability into cloud native environments. This is where AIOps delivers significant value.

Today’s cloud environments are characterized by complexity, scale and dynamism. They can feature thousands of digital services and billions of dependencies capable of changing in milliseconds. They also produce an immense volume and variety of data. AIOps can help organizations understand every change and dependency in real time, learn and detect new problems as they appear and crunch all that data to drive precise auto-remediation workflows. This is the kind of intelligent automation that can really take the pressure off DevOps teams. It frees them up to work on business-centric digital transformation projects by continuously analyzing and adjusting multicloud environments without the need for manual intervention.

Selling the AIOps Vision

Most IT and business leaders should be able to get on board with such a vision. However, we know first-hand the cultural challenges that can appear once you begin down this road to “automation everywhere.” Some DevOps practitioners are understandably concerned that it could be the end of their role as they know it. In reality, it’s an important evolution that will see the “Ops” part of a team finally capable of devoting all its attention to important projects, knowing that “keeping the lights on” tasks are safely automated. For the “Dev” side, it means a release from endless code review and debugging cycles to drive further productivity improvements.

Still, getting there can be a challenge, given the entrenched mindset that DevOps teams must be ready to jump on any operational alerts 24/7. Organizations must instead create a culture where applications are built to be self-healing, self-orchestrated and autonomous from the very start. As we’ve seen at Dynatrace, once the business and DevOps teams realize the huge upsides to reinventing ITOps along these lines, there should be less resistance to change.

Getting There in Three Stages

So how did we get to an AIOps-powered autonomous cloud? Our experience can be broken down into three core stages:

  • Stage 1: Speed. The first step is all about automating delivery, to speed up time to market and adopting an infrastructure-as-code approach where Ops works more like Dev. The ability to gain visibility into deployments across development, staging and production is crucial to success during this stage of the journey.
  • Stage 2: Stability. Against the backdrop of a product with a growing number of users and features to support, Stage 2 was all about driving differentiation through improved quality and stability. This was where we really started to embrace the potential of an autonomous cloud, with developers automating operational tasks like runbook execution to solve infrastructure challenges. Self-healing and auto-remediation also came into play, enabling us to drive a 93% reduction in production problems affecting end users. Culturally, you should now be in a place where no traditional operational tasks are seen as necessary to deploy and operate software.
  • Stage 3: Scale. The third and final stage is where it’s time to roll out this ‘automation everywhere’ approach across the business. We did this by extending AIOps out from the team working on our core platform to those working on billing, licensing, customer experience, proactive customer support and other services. Whatever solution you’re using to support an autonomous cloud needs to be so easy and scalable that your engineers want to use it instead of their own toolchains — an automatic, all-in-one self-service platform.

At no point in living memory has digital transformation mattered more to the future success of any organization. By embracing the shift towards an autonomous cloud, it’s possible to get development, operations and business functions working seamlessly together to respond faster with innovative new experiences to meet changing customer demands.

Lead image via Pixabay.

A newsletter digest of the week’s most important stories & analyses.