Failed Executions on writing to DB

Looking for some help on diagnosing random ‘Failed Executions’ on writing info to MSSQL. Systems have been installed for several years now and never paid much attention to the failures before. New data has become more critical now so need to resolve the issue. Haven’t found any data formats out of the ordinary. Have about 100 transactions configured, 50 are machine status change driven and the other 50 are driven by an hourly trigger. Also have substantial number of queries for status and data happening concurrently for HMI info at numerous clients. I have about a 3% failure rate that needs to be reduced. Both Gateway and MSSQL are located on the same machine as well.
Other option, is there a way I can ‘retrigger’ the write upon a failure?