In a CI pipeline running on a pull request, the team wants to confirm an Automation Bundle's configuration is syntactically valid and all referenced files exist, WITHOUT connecting to or changing any Databricks workspace. Which Databricks CLI command is appropriate for this CI step?
Answer & explanation
Correct answer: A
WHY A: databricks bundle validate performs local checks — syntax, schema compliance, variable and file references — without contacting a workspace, exactly what a PR-triggered CI check needs. WHY NOT B: deploy actually creates or updates workspace resources; it does not merely simulate, even on a pull request. WHY NOT C: bundle plan is a real command, but it contacts the target workspace to compute a Terraform-style diff of pending changes, so it violates the 'without connecting to a workspace' requirement that validate meets locally. WHY NOT D: bundle apply is not a bundle subcommand — the apply step is performed by deploy, not by a local validator. WHY NOT E: bundle run triggers an already-deployed job; it is not a configuration validator and has no dry-run validation mode.