Why are some write requests throttled whereas the average consumed write capacity is below the provisioned write capacity?

Question

I see the following graphs in one of my DynamoDB table:

Between 2/15 9:30 and 2/15 10:00 some write requests were throttled while the average consumed write capacity was below the provisioned write capacity. How comes?

Answer 1

It doesn't have to be a hot key or partitioning issue where the total provisioned write capacity is spread across the partitions.

You have a 5 minute average chart there. Technically you could have a case where you do have zero write activity for a 4:50 minute period and then have 10 seconds of ridiculous amounts of writes. These will most likely exceed you provisioned writes while the average still is below the limit.

Answer 2

Perhaps your partition key badly distributes the workload between the items. Or you may be doing multiple writes into the same items. As mentioned here:

For example, if a table has a very small number of heavily accessed partition key values, possibly even a single very heavily used partition key value, request traffic is concentrated on a small number of partitions – potentially only one partition. If the workload is heavily unbalanced, meaning that it is disproportionately focused on one or a few partitions, the requests will not achieve the overall provisioned throughput level.