When the system consumes the message in SQS, it is ready to update a record, the result of the update fails to report an error, and then it is always reported error during repeated consumption (this SQL statement should be executed very quickly, just update a field according to the primary key id). As a result, the SQS is pressed a lot of messages and cannot be consumed until the service is restarted, and the message in the queue is consumed. At this point the database is normal, other data queries and inserts are normal. The error message is as follows

1. 1. 1. The error may exist in com/test/repository/ProductRepository.java (mapper)
      2. The error may involve com.test.repository.ProductRepository.__update_by_id_ignore_null-Inline
      3. The error occurred while setting parameters
      4. SQL: update product_sales SET end_date=? where id = ?
      5. Cause: java.sql.SQLNonTransientConnectionException: (conn=581116) Communications link failure with primary host rds-other.production.net:3306. Connection timed out
         ; (conn=581116) Communications link failure with primary host rds-other.production.net:3306. Connection timed out; nested exception is java.sql.SQLNonTransientConnectionException: (conn=581116) Communications link failure with primary host rds-other.production.net:3306. Connection timed out

See the file for complete error messages.