DevOps Isn’t Dead, but It’s Not in Great Health Either

SEATTLE — Co-hosted at Open Source Summit (OSSummit) North America, the Continuous Delivery Foundation (CDF) reported at cdCon in its fifth annual State of CI/CD Report that while 83% of developers are actively engaged in DevOps, there’s been a troubling increase in the proportion of low performers in deployment metrics.

So what does that mean? DevOps is all about making it easier for developers and system administrators to complete software work quickly and efficiently. That’s not been happening.

Instead, out of the tens of thousands of developers surveyed by SlashData’s Developer Nation, a mere 14% can get code into production in a single day. That’s about the same as we saw when SlashData started asking this question in the third quarter of 2020. As for deploying code multiple times per day, it’s actually shrunk to  9%, from 11% in 2020.

Only 11% of DevOps users reported being able to restore service in under an hour.

Isn’t that the whole point of  Continuous Integration/Continuous Delivery (CI/CD)? Why, yes. Yes, it is.

Another reason we use DevOps is that when bad things happen — and they always do, usually right when the boss is looking over your shoulder — you can get things back up and working quickly. Alas, that’s not the case.

Only 11% of DevOps users reported being able to restore service in under an hour. OK, that’s hard. But, it’s not any better than where we were a few years ago. Worse still, these days, 41% of users report taking more than a week to restore service. In 2020, 34% could get things back up and running in just over a week.

So, what’s the problem? The report’s authors speculate that “It may be that the ubiquity of DevOps practices has allowed developers and organizations to increase the complexity of projects they are involved in, counteracting the benefits to development velocity. In other words, DevOps practices have likely made the development velocity of complex projects comparable to simpler projects without DevOps practices.”

They make another point: “As DevOps matures, developers go from exploring the space to focusing only on the technologies they find most useful. However, usefulness does not always directly correlate to deployment performance.”

That’s a disconnect that needs to be addressed. Just because you like, say, Jenkins, but all the teams that use GitHub Actions are more productive, spells out that you need to reconsider what DevOps tools you’re using.

Using multiple CI/CD tools, by the way, the report states, is a mistake. “Deployment performance is worse when using multiple CI/CD tools of the same form. That’s because, the researchers theorize, because of interoperability challenges.

Another related problem is that while deployments tend to go faster for people who use more DevOps tools, they also add more mental load to their work. A particularly troublesome example of this is alarm fatigue. When one program after another constantly chirps at you with one alert after another, it’s all too easy to stop paying attention and let real problems slide into your production pipeline.

With that said, don’t throw the productivity baby out with the CI/CD bathwater. Developers who do master multiple tools tend to get more done than their counterparts who don’t have many programs in their toolbox.

Properly used — note I say properly — the use of CI/CD pipelines goes hand-in-hand with better deployment performance across all DevOps Research and Assessment (DORA)  metrics. Used poorly; it’s another story.

I also found it disturbing that there’s been a slight decline in CI/CD usage. In Q3 2023, 33% of developers were using CI to build and test automatically their code changes. In Q1 2024, it was down to 29% of developers. Similarly, 29% in Q3 2023 used CD to automate their code deployments. In Q1 2024. It was 29%.

What’s going on here?

I think it’s high time that companies start asking themselves what the right blend is. They need to use DevOps and CI/CD tools specifically to get the most good from them. They are helpful. No one but Luddites doubts that. But, they don’t appear to be as helpful as they should be.

Perhaps the CIF and other vendor-neutral DevOps organizations should take a long, hard look at how we’re using these programs. Something appears to be amiss here.