What is DevOps? Unpacking the rise of an IT cultural revolution

Combining operations and development to deliver continuous software improvement can reduce complexity and improve application output. Learn more about DevOps and best practices to achieve it at scale. 

What is DevOps?

DevOps is a collection of flexible practices and processes organizations use to create and deliver applications and services by aligning and coordinating software development with IT perations. As DevOps pioneer Patrick Debois first described it in 2009, DevOps is not a specific technology, but a tactical approach. By working together, development and operations teams can eliminate roadblocks and focus on improving how they create, deploy, and continuously monitor software.

The shift to DevOps is critical for organizations to support the ever-accelerating development speeds that both customers and internal stakeholders demand. With the help of cloud-native technologies, open-source solutions, and agile APIs, teams can now deliver and maintain code more efficiently than ever. Combining development with operations and the processes that support them enables organizations to keep pace with the speed of development.

DevOps infinity loop

The origin of DevOps

DevOps got its start in 2008 with developers Andrew Clay and Patrick Debois. Looking to overcome common issues in agile development — such as decreased collaboration as project timelines expand, and the negative impacts of incremental delivery on long-term outcomes — the pair proposed an alternative: continuous development and delivery in a combined DevOps pipeline. The term gained traction after DevOpsDays in 2009, quickly establishing itself as a new industry buzzword.

More than a decade later, it’s clear the DevOps framework is more than just hype. In practice, its biggest benefit isn’t a simple efficiency boost, but rather a cultural shift that fundamentally changes the way companies approach every stage of the software development process.
Recently, DevOps has undergone a more in-depth evolution thanks to the work of industry experts such as Gene Kim, keynote speaker at Perform 21 and author of The DevOps Handbook and The Phoenix Project.

Why adopt DevOps?

As consulting firm Deloitte notes, technology teams are now expected to deliver projects four times faster with the same budget, and most of that budget goes toward running the business — not process or software innovation. Dynatrace’s own 2020 Global CIO Report backs up these findings: organizations spend an average of $4.8 million annually on performing manual, routine tasks just to “keep the lights on.” Siloing, the traditional separation of development and operations teams in terms of their priorities as well as the resources and insights available to them, is one major challenge that contributes to these problems.

When developers are only focused on getting code out the door, and operations teams are solely responsible for monitoring and management, there’s a blind spot in implementation — a no-man’s-land of “Who’s responsible?” In this scenario, code is kicked back to developers for revision if it doesn’t meet operational requirements, often without clear direction about what needs to happen or why. DevOps looks to eliminate these silos by creating a combined process that focuses on visibility-powered collaboration.

How does DevOps work?

To achieve this combined process, many organizations consolidate development and operations in a single team, organizing software delivery by feature rather than by job function. This approach encourages individuals to develop cross-functional skills, folding testing and even application security practices into a seamless delivery life cycle.

Implementing DevOps often goes hand in hand with continuous integration (CI), where multiple developers commit software updates to a shared repository, often many times a day. CI enables developers to discover integration issues and bugs earlier in the process, and streamline code branches and builds.

With this holistic point of view, engineers can collaborate on common processes, such as defining service-level objectives (SLOs), testing, and quality gates that everyone can implement. A common set of standards and goals can streamline agile workflows and make it possible for teams to adopt a coordinated DevOps toolset so they can automate more processes in the software delivery life cycle (SDLC).

DevOps vs BizDevOps processes

The benefits of DevOps

In practice, DevOps offers benefits not only to creating, delivering, and maintaining software, but to every process and stakeholder from early-stage proof-of-concept to digital business analytics and customer experience.

For development teams, the goal is to recognize the process of creating code as an ongoing cycle rather than a straight line. Partnering or integrating development with operations teams helps apply the principles of agile development — quick, bite-sized improvements based on priority — to the entire software life cycle. This includes initial design, proof of concept, testing, deployment, and eventual revision.

Taking this approach is especially important as both customer needs and expectations from the C-suite intensify. Development teams tasked with producing and deploying software as quickly as possible are now capable of doing just that. Meanwhile, operations teams, understandably, have concerns about the impact of rapid-fire code implementation and the changes necessary for putting that code to work, reliably and at scale.

For operations, a collaborative approach to deployment makes it possible to extend agile processes past software into platforms and infrastructure to analyze details and context of all layers within the IT environment. By applying design thinking to delivery systems, operations teams can shift their focus from managing infrastructure to delivering outstanding user experiences.

In effect, this dev-and-ops effort looks to leverage rather than limit the impact of development on operations by applying software development principles to every aspect of IT — all while maintaining operational focus on standardization and security.

How DevOps is transforming enterprises

A successful DevOps initiative is characterized by a culture of experimentation, risk, and trust — one in which constant feedback between all members is both welcomed and utilized. But culture alone isn’t enough to transform enterprise effort; teams also need the right technologies to get the job done.

As tools and technologies proliferate, a crucial capability is observability: the ability to instrument and monitor telemetry data from across the cloud-native environment. This includes metrics, logs, distributed traces, as well as data from user-experiences and the latest open-source standards to measure the health of applications and their supporting infrastructure at every stage of development.

According to a recent Gartner report, leaders should consider solutions during pre-production to maximize insight into app performance, service availability, and overall environmental health.

There’s also an emerging push to include more disciplines in the DevOps effort: “DevSecOps” teams seek to integrate security testing in delivery and deployment pipelines, while “BizDevOps” efforts strive to understand app performance from a user-experience perspective.

Although the advantages of DevOps are clear, the culture and technology shift it brings can make implementing DevOps an iterative journey.

DevOps best practices

Integrating disciplines, tools, and processes across the organization requires planning and coordination. Here are some best practices organizations can follow to make DevOps an enterprise-wide success.

Augment the DevOps pipeline with AI

Every stage of the DevOps pipeline requires some amount of analysis to drive decisions, responses, and automation.

For example, precise AI-based analysis can drive decisions about whether to release a piece of software, and once the software is in production, indicate whether the release is performing as it should. Or, during failed test runs, AI can provide the exact root cause down to the details of the underlying code, so developers can quickly address and remediate errors.

The ability to analyze data accurately and reliably, and provide definitive answers makes it possible for teams to automate processes throughout the software delivery lifecycle. Reliable, AI-driven answers are crucial for rapid incident response and auto-remediation so teams can understand the context behind a failure or error.

This artificial intelligence for IT operations (AIOps) is becoming a widespread practice, especially as organizations adopt cloud-native infrastructure.

Shift-left service-level objectives (SLOs)

To ensure development teams and SREs are aligned on the same success criteria, they should evaluate production SLOs against pre-production environments. By expanding quality assurance to include prevention, detection, and recoverability using production-level criteria, teams can deliver software that meets user requirements, reduces error rates, and increases overall reliability and resilience. More importantly, the cost of fixing errors in pre-production is much lower than in production.

A continuous quality mindset enables teams to architect the entire SDLC for testing. This means testing all layers of the lifecycle. It also means developing and maintaining reliable test data and test environments that developers, SREs, and IT operations teams can use at every stage of development and delivery.

One way to automatically evaluate pre-production SLOs is by establishing quality gates. A quality gate helps teams determine whether a service meets all pre-defined quality criteria. Quality gates take in key service level indicators (SLIs) or monitoring metrics and evaluates them against the quality criteria set. Code progresses to the next stage of the lifecycle only if the services meet or exceed the quality criteria.

Automate all DevOps processes

Automating DevOps pipelines allows for iterative, incremental software updates to be deployed faster and more frequently. It provides tighter feedback loops between Dev and Ops teams so they can spend more time innovating instead of performing manual processes.
Automation is what helps teams scale DevOps from a lighthouse project to an essential practice across the entire IT estate. DevOps automation typically rings the continuous integration, continuous delivery (CI/CD) bell, but automating these foundational processes can extend well beyond developing code. More advanced organizations seek to automate all stages of the DevOps lifecycle including infrastructure provisioning, deployment, monitoring, testing, remediation, and so on.

Adopt cloud-native architecture

To deliver on DevOps potential, speed and agility are key. The best way to deliver more and richer features faster, flexibly, and at scale is to adopt cloud-native technologies and architecture. These technologies include container-based computing solutions, such as Kubernetes, and serverless platforms-as-a-service (PaaS), such as AWS Lambda, Google Cloud Platform, and Azure Functions. In these environments, software runs in immutable containers using resources as needed, a setup that lends itself well to infrastructure patterns that can be easily orchestrated and automated.

Cloud-native technologies make it easier for teams to apply agile software development practices to infrastructure management. This includes automating key tasks such as version control, unit testing, continuous delivery, operations functions, and problem remediation.

Integrate security practices for DevSecOps

The diversity and flexibility of cloud-native technologies also makes securing applications against vulnerabilities more challenging. As mentioned earlier, integrating application security and vulnerability assessment into a DevOps workflow is a best practice that extends the benefits of AI-driven analysis and automation to securing applications.

By automating application security testing to continuously analyze applications, libraries, and code at runtime, teams can eliminate security blind spots and false-positive alerts. Adding security-related SLOs, testing, and quality gates into all phases of the delivery life cycle enables teams to cultivate a security mindset that eliminates another silo and results in more secure software.

Adopt a platform-driven approach with self-service processes

Achieving widespread DevOps success requires a platform approach that can make it easier for organizations to effect structural changes that optimize the way teams work. A key goal is to establish self-service processes for managing different types of testing, monitoring, alerting, CI/CD workflows, internal infrastructure and development environments, and public cloud infrastructure. When teams have access to reliable data and analysis, and individuals have more autonomy to rely on their own knowledge and experience, organizations can extend the value of DevOps to the whole enterprise.

With self-service access to APIs, tools, services, and support in a single platform that leverages AI and automation, teams have a single reliable source of knowledge and coordination. This enables teams to integrate and streamline their DevOps toolchains and processes so they can spend less time maintaining the infrastructure and more time innovating.

Closing the gap

What is DevOps? It’s a cultural and tactical shift that closes the gap between development efforts and operational obligations by combining teamwork with technology to streamline software delivery, standardize testing and quality gates, and automate processes and incident response. Armed with best practices and an AI-driven software intelligence platform to manage the entire DevOps tool chain, teams can maximize efficiency, minimize error rates, and deliver on continuous delivery expectations.

Watch webinar

Join us for an on-demand performance clinic, Risk Free Delivery with Dynatrace Cloud Automation Release Management, to learn how Dynatrace integrates with your progressive delivery toolset and automates the software delivery process.

Stay updated