DynamoDB Update / Throttling Despite High Throughput

Question

I see some changes in my updates on the DynamoDB table. I know that the throttle operates per second, that peaks above the allocated capacity can sometimes be absorbed, but are not guaranteed. I know that a uniform load distribution is expected, which I have not done yet.

BUT, please look at the average graphs for 1 minute from the indicators; attached. The power used is below the set volume. Where do these chokes come from? Because all the records went to a particular shard?

No batch records. Workload distribution is something that is impossible, easy to control.

Answer 1

DynamoDB , ( /) ( ).

- , , , , . , , , , . , 10 , . , , , , , .

, . , , , . DynamoDB IOPS , , IOPS , .

, , 10 100 , . AWS SDK, , , .

, , , /, , . , Memcached Redis, , . , , , .

, : DynamoDB , , HTTP-. , DynamoDB.