High availability or sharding during heavy transactions on AWS RDS

I am running my server on AWS RDS but during heavy transaction, the master slave replication crashed. The slave was rebuilt and to prevent such from happening in the future I am thinking of HA or Sharding of the database. I need help