Declarative vs scripted pipeline syntax: when to choose each?

WHAT THIS TESTS: Whether you understand the architectural trade-off between Jenkins opinionated abstraction and raw Groovy flexibility, and whether you can recommend a sensible default for a new project rather than defaulting to the most powerful tool available.

A GOOD ANSWER COVERS: First, syntax structure: Declarative Pipeline requires a top-level pipeline block and enforces a specific grammar of sections like agent, stages, and steps, while Scripted Pipeline is written in plain Groovy with no enforced structure. Second, guardrails versus freedom: Declarative limits what you can do at the top level to prevent anti-patterns, but still allows Groovy inside a script step; Scripted offers unrestricted flow control, variables, and logic. Third, limitations: Declarative has a known code-size limit inside the pipeline block that does not apply to Scripted. Fourth, selection criteria: choose Declarative for new projects because it standardizes readability, simplifies onboarding, and handles most CI/CD needs; choose Scripted only when you genuinely need complex Groovy logic, custom flow control, or you are hitting Declarative size constraints.

COMMON WRONG ANSWERS: Saying Scripted is always more powerful and therefore better for every project. Claiming Declarative cannot run any Groovy at all. Recommending Scripted for a new project simply because you know Groovy better, without considering team maintenance and readability. Asserting that the two syntaxes are just style preferences with no functional differences.

LIKELY FOLLOW-UPS: How do you handle complex logic inside a Declarative Pipeline? When would you use a Shared Library instead of switching to Scripted? How do you migrate an existing Scripted Pipeline to Declarative? What happens if your Declarative Pipeline grows beyond the code-size limit?

ONE CONCRETE EXAMPLE: Imagine a standard microservice build: checkout, unit test, build Docker image, push to registry, deploy to staging. A Declarative Pipeline uses an agent any block, stages for each phase, and built-in steps like sh and docker.build. It is readable and enforces structure. If the same project later needs dynamic stage generation based on a JSON manifest or complex exception handling across parallel branches, a Scripted Pipeline or a Declarative Pipeline with a script block becomes necessary. For the initial greenfield project, Declarative is the right starting point.