DevOps as a methodology for accelerating the development and deployment of applications has been unevenly employed for more than a decade. While there are many organizations that almost exclusively employ best DevOps practices to automate IT processes as much as possible, there are many organizations where reliance on these concepts is limited to a small number of departments within a large enterprise.
The primary reason for the lack of widespread adoption of best DevOps practices has a lot more to do with the culture of the organization than with any challenges involving a technology platform. A survey of 600 IT professionals conducted by mabl, a provider of a test automation platform, finds the biggest barriers to adopting DevOps practices are more closely related to organizational issues (82%) rather than any technology. Only 11% of respondents said their organizations have fully implemented DevOps practices. On the plus side, nearly a quarter (24%) said they were most of the way toward achieving that goal. The top two barriers to adoption are slow processes and speed of adaptation (29%), followed by budget and funding (21%), the survey finds.
Greater Adoption with Observability
There are, however, two major technology trends that could make it a lot easier for organizations to adopt best DevOps practices at a deeper level. The first is the rise of observability platforms that promise to provide IT teams with a lot more operational context than they typically have today. The second are machine learning algorithms that are at the core of IT service management (ITSM) platforms infused with artificial intelligence (AI), otherwise known as AIOps. Observability platforms are likely to have a more widespread impact in the short term, but over time the role AI will play in advancing DevOps may have a more profound impact.
Observability has always been a core tenet of any best DevOps practice. Achieving that goal has always been more easily said than done. Most of the IT management tools relied on today are optimized for specific platforms and applications. Whenever there is an issue, IT teams typically convene a “war room” where they try to uncover the root cause of, for example, a performance issue by comparing reports and the various dashboards generated by all the tools they employ. It can take weeks to discover an issue that often only takes a few minutes to fix once the source of the problem is identified.
A new generation of observability platforms is promising to reduce the mean time to discovering IT issues by aggregating all the data generated by the IT environment within a single tool designed to correlate events that makes it easier to identify anomalous behavior. Armed with those insights it becomes a lot simpler for IT teams to drill down into the reports and alerts generated in a way that ultimately services to resolve issues faster.
Also read: AIOps Trends & Benefits for 2021
Employing APM Platforms
There are not only a larger number of startup vendors providing observability platforms, but just about every provider of an application performance management (APM) platform is also repositioning its offering as an observability platform. At the core of these efforts is open source agent software that is substantially reducing the total cost of instrumenting IT infrastructure and applications.
Previously, usage of APM platforms was generally limited to mission-critical applications because deploying agent software to collect data incurred additional costs. In the meantime, the process of deploying agent software is slowly becoming more automated. In addition, the cost of acquiring an observability platform can be mitigated by rationalizing many of the existing IT management tools that an organization may no longer require once the observability platform is fully operational.
The ability to collect data across the IT environment more affordably also plays a critical role in enabling AI. The data collected by these platforms is needed to train the AI model that is being deployed to automate various IT processes. The challenge is that each IT environment is unique. It takes time for the AI model to learn the environment. In some regards, it’s no different than training a new IT administrator. The difference is once the AI platform learns something it never forgets, takes a day off, or decides to leave the company because they got a better offer somewhere else.
Pairing DevOps and AIOps
Best of all, from a DevOps perspective, the opportunity to employ AI to automate IT processes at scale is tremendous. A small DevOps team will soon leverage AIOps platform to build and deploy a much larger portfolio of applications faster without having to add large numbers of additional personnel that have acquired DevOps certifications.
IT professionals that have those certifications are, of course, in high demand. Hiring and retaining DevOps specialists is a major challenge for organizations. The typical site reliability engineer (SRE) that manages the workflows around a continuous integration/continuous delivery (CI/CD) platform makes tens of thousands of dollars more than the average IT administrator. AI promises to automate many rote tasks within a DevOps process that today still require a significant amount of manual effort. A survey of 300 SREs conducted by Catchpoint, an IT monitoring platform provider, in collaboration with VMware and the DevOps Institute, suggests that the amount of toil — defined as low-level manual tasks — only declined 15% in the last year despite all the advances in automation that have been achieved.
AI and observability platforms are not, of their own accord, going to eliminate cultural hurdles that conspire to prevent organizations from fully embracing DevOps. They may, however, go a long way to reducing the size of those hurdles.