We've now got @haveibeenpwned running on @Azure SQL Hyperscale, a serverless model you can quickly scale available cores up and down on. However, we're often seeing login failures during the scaling process that lead to app exceptions, anyone else had this happen?