You probably shouldn’t use DynamoDB
Jono MacDougall

I am not sure partitioning the table on a customer id is a good choice for Dynamo. Depending on the method of customer id generation it does not necessarily guarantee a random choice (for example a sequential incrementing number would be a hot spot).

Partitioning on a more definite distribution (geo area, last name, etc) might be a better choice as a primary key. This is not really new and is a similar concept in older relational databases partitioning schemes (like Oracle for example).

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.