SQL Sentry shows the login, host, and application for queries in some areas of the product. So it can be done.

I get that this type of data would be lost in aggregation after a set number of days or weeks, but only having the detail for set time would be sufficient.

If a query is throwing an error or causing a major performance issue, I need to who ran it. In Azure SQL DB, I would have to keep an Extended Events Session running to capture that. With hundreds of databases, that creates not only performance overhead, but a maintenance issue.

We can see how this would be valuable, but at the moment, the way we sample top queries via the aggregated DMV mean this information isn't practically available to us. It's something we are going back to and reevaluating periodically however. Thank you for bringing it up.

This is something that I should find hugely beneficial!