While analyzing the performance of an Azure SQL Database with huge workload (Business-Critical service tier), I noticed the Log IO Percentage is hammered and hits 100% for considerable time periods, which as a consequence affects the overall performance. The database is being populated by several Data factory pipelines, that embody SSIS packages and stored procedures, and using INSERT/UPDATE statements extensively.
Back in on-premise world, I would change the database recovery model to Simple or Bulk-Logged, and use TABLOCK hint in my inserts, and the minimal logging is achieved (satisfying some other conditions).
Is this kind of minimal logging (TABLOCK) also applicable to Azure SQL Databases ? (I read they are in Full recovery model by default).
How to achieve minimal logging in the Azure SQL Database described above, using the same pipelines?