How to find which databases are missing a table, column or type change
Run a schema coverage scan over a target group to see exactly which databases differ, then prove the fix with a second scan.
Publicado el
After a rollout, or before one, the question is "are these databases actually the same?" Checking by hand means one query per database. A coverage scan asks them all.
Run the scan
Select your target group and run Scan coverage from the command palette or the menu. It asks first, because it opens a connection to each database to read its schema:
This opens a connection to each of the 12 databases in the selected distribution list and reads its schema.
It reads schemas only and executes nothing. You can stop it at any point and keep what was read.
Read the findings
Findings are counted against the databases actually read:
dbo.CustomerNote missing from 9 of 12
dbo.Customer.LoyaltyTier missing from 6 of 12
sales.OrderHeader.Total typed 2 ways across 12: decimal(18,2), money
That is what a half-finished rollout looks like. The same notes appear as annotations in the editor's IntelliSense, for example "missing on 6 of 12 scanned".

Fix it, then scan again
Run your fix across the group, then scan again. The first scan is the hypothesis; the second is the evidence. A scan that reports nothing means nothing was observed to differ among the databases it read, so check the count.
Read more in the schema coverage scan. To run the same check in a pipeline, see failing a build when databases drift.