At the October 2021 DevOps Enterprise Summit, several of us met to discuss the problem of the consistent misunderstanding of continuous delivery and decided to run an improvement experiment.
In the 2017 book "Accelerate", the authors observe that:
"Continuous delivery improves both delivery performance and quality, and also helps improve culture and reduce burnout and deployment pain."
This observation was based on their experience and also on the information they'd collected from years of State of DevOps survey results. We have also observed these outcomes ourselves, but only when we properly use a core set of practices. When these practices are not used and an organization misunderstands or tries to cut corners, then the improvement outcomes do not materialize and they claim that CD is a fad or not a fit for their special case. Neither is true.
By establishing a baseline, we hope to provide a clear roadmap of core abilities that must be in place. If we ask ourselves, "why can't we do these things?" and then solve those problems, we should see the benefits that other organizations report.
If you are a practitioner and agree with this goal, feel free to submit a pull request to add yourself.
We do not want to hold anyone hostage. Our goal is elevating the conversation by having a common language to begin with. You can either open an issue or submit a PR. We will promptly update.
The maintainers operate by consensus. We have a specific goal that any practice must be essential in every context to make the list. If a practice isn't there, it either doesn't meet the absolute minimum standard in every context or it conflicts with one of the current practices. Creating a GitHub issue is the best way to suggest an update to a practice.
Oh, gosh no! This is the barest of minimums.
CD requires CI and CI requires Trunk-Based Development. Start at the bottom and work your way up. Ask yourself, "why can't we do this yet?" Solving that problem is the engine of org improvement. We have not seen a single instance where CD cannot apply. It just takes longer in some situations to solve the problems. It's worth it though. It improves everything.
"If the pipeline says everything looks good, that should be enough - it forces the focus on what 'releasable' means." - Dave Farley