Hi - We have 3 Meltano jobs running in AWS MWAA containers. We are running tap-salesforce (a local variant) to target-snowflake (transferwise). Our Meltano docker image is pegged to meltano/meltano:v2.20; the tap and target are also pegged to specific versions. At 11:49 on Saturday morning NZST (Friday 23:49 UTC?) all three started failing with CRITICAL Config is missing required keys: ['select_fields_by_default'] I can see no change to the meltano.yml file used for the container image. Does anyone have any ideas about what might have happened? Edit: We have a completely separate project running in a different environment (Airflow rather than MWAA) which has also started failing at the same time. This parameter has NEVER been set in any of these environments - why has it suddenly started failing? It has always been missing...

Relevant: https://github.com/meltano/hub/pull/1512#issuecomment-1712532193

Sorry about this - we’re going to revert the change for now. Definitely check out lockfiles which will prevent this sort of thing from happening in the future!

So is the issue that we have pinned the immediate versions of the tap that we are using (a local variant of tap-salesforce@v1.0.0), and the Meltano installation (Docker image meltano/meltano:v2.20), but the former is pulling down whatever is the current base for tap-salesforce, which in this case was changed unexpcectedly. Can the variant of the tap pin itself to a particular version of the parent, to stop this happening in the future? Also, I don't really understand lock files - would it be appropriate for them to be created automatically if not there, rather than breaking existing pipelines?