ARA-C01 Question 168
Single answerSupport of DevOps/DataOps principlesA data platform team is introducing DevOps/DataOps practices for a Snowflake-based analytics environment with separate DEV, TEST, and PROD accounts. The team wants to promote database changes through automated CI/CD pipelines while minimizing deployment risk, ensuring repeatability, and allowing rapid rollback of data structure changes when a release fails. Which approach BEST supports these goals?
- A
Store all DDL scripts in a version control system, validate them in CI, and use zero-copy cloning of databases/schemas for test environments before promoting approved changes through automated deployment pipelines
- B
Allow developers to make direct schema changes in PROD because Snowflake Time Travel can recover dropped objects if a release fails
- C
Export database objects from DEV weekly, manually compare them to PROD, and apply differences during a scheduled maintenance window
- D
Use a single Snowflake account for DEV, TEST, and PROD to simplify object promotion, and separate environments only by warehouse names
Show answer and explanation
Correct answer: A
Explanation
The best answer is to treat Snowflake changes as code and promote them through automated pipelines, using version control and environment validation. This directly supports DevOps/DataOps principles such as reproducibility, auditability, peer review, automation, and lower-risk deployments. Snowflake capabilities that reinforce this model include zero-copy cloning for fast environment creation and testing, role-based access control for separation of duties, and native SQL-based object management that can be integrated with CI/CD tooling. While Time Travel and Fail-safe provide recovery capabilities, they do not replace disciplined release engineering. Snowflake best practices generally emphasize automation, environment isolation, and use of cloning for testing and validation in support of modern data delivery workflows.
- A. Correct.
Correct. This approach aligns with DevOps/DataOps best practices: infrastructure and database changes are defined as code, stored in version control, validated automatically, and promoted through CI/CD. Using zero-copy cloning enables fast, low-cost creation of test environments that closely mirror source data structures and optionally data, improving release validation and repeatability. This also reduces risk because changes are tested before deployment, and cloning supports quick environment provisioning for verification and rollback workflows.
- B. Incorrect.
Incorrect. Although Time Travel can help recover certain historical data states and dropped objects within retention limits, it is not a substitute for controlled CI/CD or change management. Direct changes in PROD violate DevOps principles such as version control, peer review, repeatability, and automated promotion. This option reflects a common misconception that recovery features eliminate the need for disciplined release processes.
- C. Incorrect.
Incorrect. Manual export, comparison, and deployment introduce inconsistency, delay, and human error, which are exactly what DevOps/DataOps practices aim to reduce. Weekly exports also do not support rapid iteration or reliable rollback. Snowflake environments are better managed with automated, source-controlled deployment processes rather than ad hoc manual promotion.
- D. Incorrect.
Incorrect. Using one account for all environments weakens isolation and increases the risk of accidental cross-environment impact. Warehouse-level separation does not isolate database objects, roles, policies, or account-level configuration adequately for enterprise SDLC practices. Snowflake commonly supports environment separation through different accounts, databases, and role-based controls depending on organizational requirements.