Feature Request: Option for Elementary to fail silently instead of failing the dbt run #1991
Description
Is your feature request related to a problem? Please describe.
In our production dbt environment, we run thousands of models daily with Elementary enabled. Currently, if Elementary encounters an error during execution, it can cause failures or interruptions in the workflow. This is problematic because we want monitoring and metadata collection, but we cannot risk the monitoring tool blocking or failing the actual model runs.
Describe the solution you'd like
Introduce a configuration option (e.g., fail_silently: true) that ensures Elementary logs/reporting continues where possible but does not raise errors that cause the run to fail. If an error occurs in Elementary, it should gracefully skip/ignore the failure and continue execution of the rest of the dbt run.
Describe alternatives you've considered
Running Elementary separately as a dedicated job, but this makes it more difficult to tie metadata checks directly to specific runs.
Wrapping Elementary calls in external error-handling scripts, which adds maintenance overhead and feels like a workaround.
Additional context
Fail-silently functionality would make Elementary more resilient for large-scale production environments and reduce the operational risk of enabling monitoring in dbt pipelines.
Would you be willing to contribute this feature?
Yes, I’d be open to contributing if maintainers can provide guidance on where best to integrate this option.