Retries, branching, and looping in a Lakeflow Jobs pipeline

Lakeflow Jobs let you orchestrate multi-task pipelines with control flow features like task retries, conditional branching via 'Run if' dependencies, and the If/else and For each task types. These constructs let you handle failures gracefully and dynamically execute paths based on upstream task outcomes or a list of inputs.

Must-know

• Each task can define a retry policy (number of retries and interval in seconds) so transient failures auto-retry before the task is marked failed.
• The 'Run if dependencies' setting on a task controls execution based on upstream task status: All succeeded (default), At least one succeeded, None failed, All done, At least one failed, or All failed.
• The If/else condition task type evaluates a boolean expression (comparing task values, parameters, etc.) and routes execution to different downstream tasks based on the result.
• The For each task type loops over a list of inputs, running a nested task once per item, with a configurable concurrency limit for parallel iterations.
• Task values (set via dbutils.jobs.taskValues) can be passed between tasks and referenced in downstream conditions or For each input lists to drive dynamic control flow.
• A job run is marked failed if any task fails and no downstream task's Run if condition allows the job to continue, so branching logic must account for overall job success criteria.

Coverage checked against the published exam guide on Jul 26, 2026.

These are independent practice questions, written against this certification's published exam guide. They are not the certification vendor's own questions, and not the real exam.