The following is the RCA from Microsoft regarding this issue:
What happened:
Between June 5, 2025, 18:15 UTC and June 6, 2025, 01:29:14 UTC, the SQL Managed Instance pearlsqlmi experienced intermittent unavailability and connectivity issues.
Cause:
This was caused by multiple failovers triggered by a rare race condition during the rotation of the encryption protector on databases that contain memory-optimized tables.
Mitigation
Once the root cause was identified, we implemented manual mitigation steps to restore stability. Monitoring and alerting systems remain active to detect any recurrence.
This issue is highly unlikely to reoccur and is only triggered during encryption protector changes.
Next steps
Product team is currently validating a permanent fix, which will be included in an upcoming release.
In the meantime, we recommend avoiding encryption protector changes during peak usage hours to minimize potential impact, since during that period you have a slightly higher chance of this particular race condition happening.
