Overview

Released on August 22, 2019, the 2019 Accelerate: State of DevOps Report drew on survey data from approximately 31,000 IT professionals worldwide to measure the state of software delivery performance across the industry. Nicole Forsgren, researcher and strategist at Google Cloud and a co-author of the report, presented the highlights at the Open Source Summit in San Diego.

“When I’m talking about DevOps, I’m talking about software development, software delivery, and I’m also talking about availability,” Forsgren said during her keynote.

The Four Key Metrics

The report measures software delivery performance using several core metrics that have become a de facto standard for evaluating DevOps maturity:

• Lead time for changes: The elapsed time from a code commit to that code running in production.
• Deployment frequency: How often a team successfully deploys code to production.
• Time to restore service: How quickly a team can recover from a production incident or outage.
• Change failure rate: The percentage of deployments that cause a failure requiring remediation (referenced in the broader report methodology, though not highlighted in the keynote summary).

For elite-performing organizations, the 2019 report found:

• Lead time for changes was less than one day.
• Deployment frequency reached multiple deployments per day.
• Time to restore service was less than one hour.

Elite Performers Nearly Tripled Year-Over-Year

One of the most notable findings in the 2019 edition was the significant growth in the proportion of teams classified as elite performers. Forsgren highlighted the shift directly during her keynote:

“Check out those elite performers last year; 7% of our survey sample were elite performers, optimizing on speed and stability. This year, check it out—the elite level almost tripled and hit 20%.”

Nicole Forsgren, Google Cloud

At the same time, the share of low performers—teams with a deployment frequency of between once per month and once every six months—declined from 15% to 12%.

“That means that more people are getting better,” Forsgren said, framing the shift as evidence that DevOps practices are maturing across the broader industry, not just at a handful of large technology companies.

Faster Code Is More Stable Code

A counterintuitive but consistent finding in the Accelerate research is that deploying more frequently is associated with greater system stability, not less. Forsgren explained the mechanism in practical terms: when teams delay releases, code accumulates. Larger, less frequent deployments increase the complexity of each change and expand what she calls the “blast radius”—the potential scope of impact when something goes wrong.

“When we introduce delays, we intend to create a more stable system, but we’ve never seen that. It actually creates instability.”

Nicole Forsgren, Google Cloud

This finding argues against the common instinct to slow down release cycles as a risk-reduction measure. Smaller, more frequent deployments reduce the blast radius of any individual change and make it far easier to isolate and fix problems quickly.

How to Improve DevOps Performance

The report identifies several practical levers that organizations can pull to improve their DevOps capabilities. Forsgren discussed two in particular during her keynote.

Addressing Technical Debt

Technical debt—code that is poorly maintained, disorganized, or inadequately documented—is a direct barrier to engineering productivity. Forsgren defined productivity in this context as “the ability to get complex, time-consuming tasks completed with minimal distraction and interruptions,” and was explicit that “technical debt reduces productivity.”

The report points to a loosely coupled architecture combined with continuous monitoring as practical means of keeping technical debt in check. A loosely coupled architecture allows teams to deploy and update individual services independently, reducing the coordination overhead and risk associated with tightly coupled systems.

Building a Culture of Trust and Psychological Safety

Beyond technical practices, organizational culture plays a significant role in DevOps performance. The report found that giving developers visibility into the organization’s goals—and fostering an environment where people feel safe to take risks and raise concerns—has measurable positive effects.

“Trust and psychological safety have a positive impact on software delivery performance, organizational performance and productivity.”

Nicole Forsgren, Google Cloud

This aligns with a long-standing theme in the Accelerate research series: high-performing DevOps teams are not simply a product of better tooling. They are a product of better communication, clearer goals, and a culture that treats experimentation and learning from failure as normal parts of the development process.

Key Takeaways for DevOps Teams

• Measure what matters: Lead time, deployment frequency, and time to restore are proven indicators of DevOps maturity and should be tracked consistently.
• Deploy smaller, deploy more often: Batching changes does not reduce risk—it amplifies it. Smaller, more frequent releases are demonstrably safer.
• Invest in architecture: Loosely coupled systems reduce both the cost of change and the accumulation of technical debt.
• Culture is not a soft metric: Psychological safety and organizational trust have a direct, measurable effect on software delivery performance.
• The industry is improving: The jump from 7% to 20% elite performers in a single year suggests that the practices advocated by the DevOps community are reaching mainstream adoption.

Jenkins Turns 15: From a Broken Build to 15.8 Million Developers

In 2004, Kohsuke Kawaguchi was a software engineer at Sun Microsystems with a habit of accidentally breaking builds. Frustrated by colleague callbacks, he asked himself a simple question: could a program catch his mistakes before his colleagues did? That question led him to create the automation server now known as Jenkins — and, in many ways, to help define the modern era of continuous integration.

Fifteen years later, Kawaguchi serves as Chief Scientist at CloudBees, the company that provides commercial support, services, and a growing product portfolio built around Jenkins. At DevOps World | Jenkins World 2019, held in San Francisco, CloudBees marked the anniversary while also making a series of significant product announcements. According to data provided by CloudBees at the event, approximately 15.8 million developers worldwide use Jenkins as part of their DevOps workflows.

What Is Jenkins?

Jenkins is an open-source automation server primarily used to implement continuous integration (CI) and continuous delivery (CD) pipelines. It enables development teams to automatically build, test, and deploy code changes, reducing manual effort and catching integration errors early. Its extensibility through a large plugin ecosystem has made it one of the most widely adopted CI/CD tools in the industry.

CloudBees Flow 9.1 and Accelerator 11.1

Jenkins is foundational to CloudBees, but the company has been expanding well beyond it. At DevOps World, CloudBees announced the 9.1 release of CloudBees Flow, its Application Release Orchestration (ARO) platform. Flow came to CloudBees through its April 2019 acquisition of Electric Cloud, which had developed the product to coordinate complex, multi-team software releases across environments and pipelines.

The same acquisition brought CloudBees Accelerator 11.1, a rebranded version of Electric Cloud’s build and test acceleration product. According to CloudBees CEO and co-founder Sacha Labourey, the tool is designed to compress long build and test cycles:

“The accelerator makes it possible to accelerate build and test from hours to minutes.”

— Sacha Labourey, Co-founder and CEO, CloudBees

Feature Flags via the Rollout Acquisition

In June 2019, CloudBees also acquired Rollout, a DevOps vendor specialising in feature flag management. Feature flags (also called feature toggles) allow teams to enable or disable specific features in production without redeploying code, enabling controlled rollouts, A/B testing, and rapid rollback. The Rollout technology was rebranded and integrated into the broader CloudBees portfolio.

Labourey described the value of the capability in terms of experimentation speed:

“With Rollout, you get to experiment, measure and observe the behavior of what you want to test. And once you’re done, you can switch it off in seconds. Remember, time is acceleration, so it’s important for you to conclude your experiments ASAP.”

— Sacha Labourey, Co-founder and CEO, CloudBees

Software Delivery Management: The Big Reveal

The headline announcement at DevOps World 2019 came on August 14 with the preview of CloudBees’ Software Delivery Management (SDM) platform. The platform aims to give both developers and engineering leadership a unified, contextualised view of the state of software delivery across an organisation — aggregating data from the many disparate tools that modern development teams use every day.

Product Hub: Centralised Visibility

At the core of the SDM platform is a component called Product Hub, described by Ben Williams, Vice President of Product Management at CloudBees, as the central orchestration point for all software delivery activity:

“The Product Hub is the heart of your software delivery efforts. It’s where all of your software delivery activity is visible and orchestrated from.”

— Ben Williams, VP of Product Management, CloudBees

The dashboard ingests data from tools including Jenkins, Jira, GitHub, and observability and monitoring platforms, then presents that data in a single place with contextual organisation. Williams described the goal as consolidating a fragmented toolchain view into a coherent delivery picture for stakeholders at every level.

Efficiency Dashboard and Policy Engine

Beyond visibility, the SDM platform includes an efficiency dashboard paired with a policy engine. The policy engine allows organisations to define and enforce process rules — for example, tracking how long a work item remained idle before progressing. This kind of flow metric is closely aligned with concepts from value stream management, helping teams identify bottlenecks and waste in their delivery pipelines.

Extensibility via an App Framework

The SDM platform is designed to connect to a wide range of DevOps and CI/CD tooling through an open app framework. According to Williams:

“The app framework allows users and partners to be able to create custom integrations, to be able to contribute policies and rules to extend the data model, and to build custom functionality.”

— Ben Williams, VP of Product Management, CloudBees

At the time of the announcement, the SDM platform was available in limited preview, with broader availability and additional integrations planned for subsequent releases.

Key Takeaways

• Jenkins at 15: The open-source CI/CD server, created in 2004 by Kohsuke Kawaguchi, had reached approximately 15.8 million developers worldwide by mid-2019.
• CloudBees Flow 9.1: Application Release Orchestration platform acquired from Electric Cloud, now integrated into the CloudBees portfolio.
• CloudBees Accelerator 11.1: Build and test acceleration tool, also from the Electric Cloud acquisition, aimed at reducing cycle times from hours to minutes.
• Rollout integration: Feature flag management capability acquired in June 2019, enabling controlled feature rollouts and rapid experimentation.
• Software Delivery Management (SDM): A new cloud-based platform previewed at DevOps World, centred on a Product Hub that aggregates delivery data, an efficiency dashboard, a policy engine, and an extensible app framework.

Original reporting by ITPro Today. Announcements were made at DevOps World | Jenkins World, San Francisco, August 2019.