you are correct about labkey.xml, but I think there's a second thing happening here:

I defined two different external schemas in labkey. they both use the 'labkey' data source (which looks like it is the postgres db labkey2), then I pointed it to the 'Audit' schema and gave it the name 'audit'. I then did exactly the same thing, but named it 'audit_db'. links below:

https://xnight.primate.wisc.edu:8443/labkey/query/WNPRC/EHR/executeQuery.view?schemaName=audit&query.queryName=auditlog

https://xnight.primate.wisc.edu:8443/labkey/query/WNPRC/EHR/executeQuery.view?schemaName=audit_db&query.queryName=auditlog

these return a different number of rows. the one called 'audit' returns a couple thousand, while the one called 'audit_db' returns 5 million. does the name 'audit' collide with something in labkey? does the 'audit' db somehow get labkey's default filter-by-container behavior? at least as it is exposed in the schema browser, the audit table provided by labkey is labeled 'auditLog'.

by giving the external schema a different name, it appears to work properly now. i understand the rationale behind avoiding exposing the underlying postgres data, but since this should be read-only I dont see a whole lot of harm here.