Do you use non-identifying job parameters?

[fmbenhassine]

Do you use non-identifying job parameters? If yes, please elaborate on your use case by adding a comment on this poll.

[christophejan]

I use non-identifying job parameters to switch to a degraded processing of some data or a specific part of the processing.

Possible causes include service maintenance, an incident, inconsistent data, or an unexpected breaking change.

This is particularly helpful when nominal processing fails, leaving half-processed data. This is a common issue, and we need to complete the processing quickly and minimise further manual intervention.

Using non-identifying job parameters to control such specific behaviour seems the most appropriate approach as:

• They don’t identify the processed data. It’s not a different job or dataset, and in case of a restart, it should be able to resume the job instance from where it left off.
• They are specific to a job execution. Using instead Spring environment seems inappropriate as the same environment may be used sometimes for different job executions.