Is ok to have multiple Serverless replications in DMS using the same source endpoint?

Yes, it is generally acceptable to have multiple serverless replications in AWS Database Migration Service (DMS) using the same source endpoint. This approach can be beneficial in certain scenarios, particularly when dealing with large-scale migrations or when you need to partition your workload.

AWS DMS Serverless has a limitation of 100GB allocated storage per replication. If your migration requires more memory than this, perhaps due to long-running transactions or caching needs, it's recommended to partition your workload into separate serverless replications. You can divide your workload based on different criteria, such as by table or by specific requirements. For instance, you might choose to put all replication involving LOBs (Large Objects) into a separate serverless replication.

By using multiple serverless replications with the same source endpoint, you can effectively manage larger migrations while staying within the constraints of the service. This approach allows you to optimize your migration strategy and potentially improve performance by distributing the workload across multiple replications.

However, it's important to carefully plan and manage these multiple replications to ensure they don't interfere with each other or overload the source system. Always consider the impact on your source database and monitor the process closely to maintain data consistency and performance.

Sources
AWS DMS Serverless limitations - AWS Database Migration Service