QuestionQ79

VerifiedDeployment and Orchestration of ML Workflows

A company implemented a data-ingestion pipeline for sales transactions from its ecommerce website. The company uses Amazon Data Firehose to ingest data into Amazon OpenSearch Service. The Firehose stream’s buffer interval is set to 60 seconds. An OpenSearch linear model produces real-time sales forecasts from the data and displays them in an OpenSearch dashboard.

The company must optimize the data-ingestion pipeline to provide sub-second latency for the real-time dashboard.

Which architectural change will meet these requirements?

Explanation

For an Amazon OpenSearch Service destination, Amazon Data Firehose can buffer incoming data for an interval from 0 to 900 seconds before delivery. Setting the interval to zero eliminates the 60-second buffering delay; tuning PutRecordBatch batch size also minimizes producer-side batching delay, enabling the lowest-latency option among the choices. Amazon Data Firehose API Reference: AmazonopensearchserviceBufferingHints

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!