Exam Bundle

This signal belongs to the adjacent case “A new bundle deployment should not reach production unless tests and validation pass,” whose remedy is Make bundle deployment depend on automated test and bundle-validation stages. The observed result for the present requirement is The test update fails and no invalid silver records are published.

A bad code change causes the fixture comparison to fail pertains to a different control (Write an `assertDataFrameEqual` unit test with expected output) and a different need (A pure PySpark transformation must be validated on fixed input fixtures). It cannot establish that Exercise a negative fixture against the strict expectation met the stated goal.

That observation is a valid indicator for “An output schema must preserve expected column types despite source changes” when using Use `assertSchemaEqual` in CI against an explicit expected schema. The scenario instead tests “A quality expectation should stop malformed data from being published”; the relevant result is The test update fails and no invalid silver records are published.

Make bundle deployment depend on automated test and bundle-validation stages is appropriate when “A new bundle deployment should not reach production unless tests and validation pass”; it would be verified by The pipeline blocks deployment after an intentionally failing test. It does not directly satisfy “A production target must not use the developer's personal credentials,” whose relevant acceptance evidence is Deployment records show the service principal as the acting identity.

Run `databricks bundle deploy -t prod` under approved credentials is appropriate when “A CI job should deploy a reviewed release into production with the correct target”; it would be verified by The target workspace shows the deployed resource definitions. It does not directly satisfy “A production target must not use the developer's personal credentials,” whose relevant acceptance evidence is Deployment records show the service principal as the acting identity.

Reference a secret scope or governed connection in the task configuration is appropriate when “A secret used in production must not be stored in the bundle YAML”; it would be verified by No secret value appears in source YAML or deployment logs. It does not directly satisfy “A production target must not use the developer's personal credentials,” whose relevant acceptance evidence is Deployment records show the service principal as the acting identity.

Feature changes run without modifying the production branch would demonstrate the separate action Use an isolated Git branch or workspace Git folder checkout, which addresses “A developer needs to test a feature branch without modifying production code.” It does not verify the requirement “A production target must not use the developer's personal credentials” following Configure an authorized service principal for production deployment.

No secret value appears in source YAML or deployment logs pertains to a different control (Reference a secret scope or governed connection in the task configuration) and a different need (A secret used in production must not be stored in the bundle YAML). It cannot establish that Configure an authorized service principal for production deployment met the stated goal.

This signal belongs to the adjacent case “A new bundle deployment should not reach production unless tests and validation pass,” whose remedy is Make bundle deployment depend on automated test and bundle-validation stages. The observed result for the present requirement is Deployment records show the service principal as the acting identity.

This action solves a different issue: “Operations need to know exactly which code and settings produced each pipeline run.” Its normal signal is Run evidence references the approved code revision and deployment target. Here the requested behavior is “A release needs environment-specific source paths without editing source code,” so that signal is not sufficient.

The alternative Use an isolated Git branch or workspace Git folder checkout has a distinct purpose—A developer needs to test a feature branch without modifying production code. It would not establish the required outcome (A release needs environment-specific source paths without editing source code); the necessary verification concerns The resolved deployment uses the correct environment source path.

Configure an authorized service principal for production deployment is appropriate when “A production target must not use the developer's personal credentials”; it would be verified by Deployment records show the service principal as the acting identity. It does not directly satisfy “A release needs environment-specific source paths without editing source code,” whose relevant acceptance evidence is The resolved deployment uses the correct environment source path.

New runs use the previous code version while bad data is handled explicitly would demonstrate the separate action Redeploy the previously validated artifact and repair affected data separately, which addresses “A release introduced bad schema output and must be reverted quickly.” It does not verify the requirement “A release needs environment-specific source paths without editing source code” following Define per-target variables in the bundle configuration.

That observation is a valid indicator for “A new team member should deploy to development but not production” when using Separate target-specific deployment permissions by identity. The scenario instead tests “A release needs environment-specific source paths without editing source code”; the relevant result is The resolved deployment uses the correct environment source path.

Feature changes run without modifying the production branch pertains to a different control (Use an isolated Git branch or workspace Git folder checkout) and a different need (A developer needs to test a feature branch without modifying production code). It cannot establish that Define per-target variables in the bundle configuration met the stated goal.

Define per-target variables in the bundle configuration belongs to the separate case “A release needs environment-specific source paths without editing source code,” where The resolved deployment uses the correct environment source path would be expected. That differs from the condition in this stem: “A job must receive the chosen fiscal month from CI pipeline inputs.”

Start a repair run with the approved parameter override belongs to the separate case “A failed job needs a corrected processing date to repair the affected period,” where The repaired task processes the requested date and records that parameter would be expected. That differs from the condition in this stem: “A job must receive the chosen fiscal month from CI pipeline inputs.”

The alternative Inspect driver logs and the job environment dependency list has a distinct purpose—A Python task raises ImportError only on the production job compute. It would not establish the required outcome (A job must receive the chosen fiscal month from CI pipeline inputs); the necessary verification concerns Run details report the expected fiscal-month value.

This signal belongs to the adjacent case “An unwanted test resource from a bundle should be deleted under controlled lifecycle,” whose remedy is Use supported bundle destroy/remove workflow for the correct target after review. The observed result for the present requirement is Run details report the expected fiscal-month value.

That observation is a valid indicator for “A script needs to start an existing job and collect its run ID” when using Call the supported Jobs run-now API or CLI command. The scenario instead tests “A job must receive the chosen fiscal month from CI pipeline inputs”; the relevant result is Run details report the expected fiscal-month value.

That observation is a valid indicator for “A secret used in production must not be stored in the bundle YAML” when using Reference a secret scope or governed connection in the task configuration. The scenario instead tests “A job must receive the chosen fiscal month from CI pipeline inputs”; the relevant result is Run details report the expected fiscal-month value.

This action solves a different issue: “A production target must not use the developer's personal credentials.” Its normal signal is Deployment records show the service principal as the acting identity. Here the requested behavior is “A secret used in production must not be stored in the bundle YAML,” so that signal is not sufficient.

The alternative Make bundle deployment depend on automated test and bundle-validation stages has a distinct purpose—A new bundle deployment should not reach production unless tests and validation pass. It would not establish the required outcome (A secret used in production must not be stored in the bundle YAML); the necessary verification concerns No secret value appears in source YAML or deployment logs.

The alternative Declare it as a task or bundle environment dependency has a distinct purpose—An external library is available to the notebook author but not to scheduled tasks. It would not establish the required outcome (A secret used in production must not be stored in the bundle YAML); the necessary verification concerns No secret value appears in source YAML or deployment logs.

This signal belongs to the adjacent case “A new bundle deployment should not reach production unless tests and validation pass,” whose remedy is Make bundle deployment depend on automated test and bundle-validation stages. The observed result for the present requirement is No secret value appears in source YAML or deployment logs.

That observation is a valid indicator for “A production change must be validated before resources are created or updated” when using Run `databricks bundle validate -t prod` under the deployment identity. The scenario instead tests “A secret used in production must not be stored in the bundle YAML”; the relevant result is No secret value appears in source YAML or deployment logs.

This signal belongs to the adjacent case “A new team member should deploy to development but not production,” whose remedy is Separate target-specific deployment permissions by identity. The observed result for the present requirement is No secret value appears in source YAML or deployment logs.

Make bundle deployment depend on automated test and bundle-validation stages belongs to the separate case “A new bundle deployment should not reach production unless tests and validation pass,” where The pipeline blocks deployment after an intentionally failing test would be expected. That differs from the condition in this stem: “A new team member should deploy to development but not production.”

The alternative Use separate target-specific source, target and checkpoint paths has a distinct purpose—Production and development deployments should never share stateful checkpoints. It would not establish the required outcome (A new team member should deploy to development but not production); the necessary verification concerns Development deploy succeeds while unauthorized production deploy is denied.

Reference a secret scope or governed connection in the task configuration belongs to the separate case “A secret used in production must not be stored in the bundle YAML,” where No secret value appears in source YAML or deployment logs would be expected. That differs from the condition in this stem: “A new team member should deploy to development but not production.”

This signal belongs to the adjacent case “A job must receive the chosen fiscal month from CI pipeline inputs,” whose remedy is Pass a validated job parameter or task parameter at run time. The observed result for the present requirement is Development deploy succeeds while unauthorized production deploy is denied.

Only required failed/downstream tasks rerun while upstream results are reused would demonstrate the separate action Use a supported repair run targeting failed tasks and dependencies, which addresses “A downstream task failed while upstream outputs are valid and should not be rerun unnecessarily.” It does not verify the requirement “A new team member should deploy to development but not production” following Separate target-specific deployment permissions by identity.

That observation is a valid indicator for “A new bundle deployment should not reach production unless tests and validation pass” when using Make bundle deployment depend on automated test and bundle-validation stages. The scenario instead tests “A new team member should deploy to development but not production”; the relevant result is Development deploy succeeds while unauthorized production deploy is denied.