In this episode Baruch (@jbaruch) hosted Kelsey Hightower (@kelseyhightower) and Melissa McKay(@melissajmckay) to discuss some of the great questions the audience asked during the "Demystifying CI/CD" webinar, they conducted a week prior. As the discussion progressed, it turned out that the answer to most of the questions is "it depends, it is really a tradeoff". Who could have thought!


Continuous delivery vs. continuous deployment. Should you automate all the way to deployment without a final, manual approval step?
Who should build pipelines and how can you trust the quality of your pipelines?
How much trust can you place on dependencies? You can’t check everything so where is the line of trust?
Imperative vs. declarative workflows. Reusability of each for different use cases? Using building blocks that encapsulate imperative actions for building declarative pipelines
Creating large blocks of workflows, with conditionals embedded in them, vs. small and containable blocks
Using mono-repositories vs. dedicated repositories for microservices or for different parts of your pipelines
Working on the production main branch vs. several feature branches
Amount of time spent on review processes like merging on a branch or merging on a pull request?
Managing changes to non-code assets like database schemas? Should they have separate pipelines? Backward compatibility and rollback considerations.
What are the merits and practicality of GitOps and using version control to drive operations?

Twitter Mentions