nerdexam
Amazon

DAS-C01 · Question #162

A company uses Amazon kinesis Data Streams to ingest and process customer behavior information from application users each day. A data analytics specialist notices that its data stream is…

The correct answer is A. Use a random partition key to ingest the records. When an Amazon Kinesis Data Stream experiences throttling due to hot shards, even if the overall data limits are not exceeded, the issue arises from an uneven distribution of records to shards. To resolve this, using a random partition key for ingesting records will distribute…

Collection

Question

A company uses Amazon kinesis Data Streams to ingest and process customer behavior information from application users each day. A data analytics specialist notices that its data stream is throttling. The specialist has turned on enhanced monitoring for the Kinesis data stream and has verified that the data stream did not exceed the data limits. The specialist discovers that there are hot shards. Which solution will resolve this issue?

Options

  • AUse a random partition key to ingest the records.
  • BIncrease the number of shards Split the size of the log records.
  • CLimit the number of records that are sent each second by the producer to match the capacity of
  • DDecrease the size of the records that are sent from the producer to match the capacity of the

How the community answered

(38 responses)
  • A
    74% (28)
  • B
    13% (5)
  • C
    5% (2)
  • D
    8% (3)

Why each option

When an Amazon Kinesis Data Stream experiences throttling due to hot shards, even if the overall data limits are not exceeded, the issue arises from an uneven distribution of records to shards. To resolve this, using a random partition key for ingesting records will distribute the data more uniformly across all shards, preventing specific shards from becoming overloaded.

AUse a random partition key to ingest the records.Correct

Hot shards occur when an uneven distribution of partition keys causes a few shards to receive a disproportionately high volume of data, leading to throttling on those specific shards. Using a random partition key ensures that records are distributed more uniformly across all available shards, thereby resolving the hot shard issue.

BIncrease the number of shards Split the size of the log records.

While increasing the number of shards can increase overall stream capacity, it does not solve the hot shard problem if the partition key distribution remains uneven, as the new shards might still become hot. Splitting log records is unrelated to shard distribution.

CLimit the number of records that are sent each second by the producer to match the capacity of

Limiting the producer's record sending rate addresses overall throttling but does not fix the root cause of hot shards, which is an uneven distribution of data to specific shards.

DDecrease the size of the records that are sent from the producer to match the capacity of the

Decreasing the size of records might allow more records per second per shard, but it does not fix the fundamental issue of uneven distribution of data to certain shards caused by a non-random partition key.

Concept tested: Kinesis Data Streams hot shard resolution

Source: https://docs.aws.amazon.com/kinesis/latest/dev/troubleshooting.html#throttling-errors

Topics

#Kinesis Data Streams#Partition Keys#Sharding#Data Distribution

Community Discussion

No community discussion yet for this question.

Sign up to join the discussion
Full DAS-C01 Practice