Behavior of Dependent Jobs when Jobs they Depend on need Automated Retry

From the official documentation[1] we can see that:

PENDING: A job that resides in the queue and isn't yet able to run due to a dependency on another job or resource. After the dependencies are satisfied, the job is moved to RUNNABLE.

I have also replicated your scenario for your better understanding:

→ lets assume B is dependent on A, where A is the failing job

→ I created a job "A" which will fail on purpose by giving the a invalid command (slee 10 in this case), I have set the 'Job attempts' to 10

→ At the very same time, I have submitted another job "B" with the dependency "A"

→ I could see that the job "B" was in 'PENDING' status till all the retries were completed

→ Once all the retries had failed for job "A", job "B" goes into 'FAILED' status as well.

So Answering your query: Q. I want to have auto-retry on the setup jobs but first want to know about the expected behaviour of the dependent data jobs -- If I put the data jobs in queue with a dependency on the setup jobs, will the data jobs fail at the FIRST failure? Or will the dependent data jobs stay in 'PENDING' state until all retry attempts are made on the setup jobs?

➢ The dependent data jobs stay in 'PENDING' state until all retry attempts are made on the setup jobs!

Reference links:

[1] Job states - https://docs.aws.amazon.com/batch/latest/userguide/job_states.html
[2] Submitting a job - https://docs.aws.amazon.com/batch/latest/userguide/submit_job.html
[3] Job dependencies - https://docs.aws.amazon.com/batch/latest/userguide/job_dependencies.html?icmpid=docs_console_unmapped