Kinesis Data Firehose
Extract, transform & load streams into multiple AWS services
TL;DR¶
- NEAR-REALTIME Delivery - (60 secs).
- Deliveries Streams into supported destinations in batches.
- Allows delivery / transformation of streams to supported destinations.
- Any HTTP endpoints
- ElasticSearch
- Amazon Redshift
- OpenSearch
- Splunk
- Datadog
Use cases¶
- Data Lake Ingestion.
- Log and Event Data Archival.
- Near Real-time Data Transformation.
- Near Real-time Analytics.
- Data Replication and Backup.
- Clickstream and User Behavior Tracking.
- IoT Data Processing.
Common Scenarios¶
- Permanent storage of stream data in S3.
- Transform and store data in different format.
- Load Stream into one of the supported destinations.
Features¶
- Support transformation of data on the fly Lambda
- Auto Scaling.
- Fully serverless.
- Resilient.
- Billing – volume through firehose.
Limitations¶
- only one consumer at the time per stream.
Performance¶
- Near real-time delivery
- batch buffer size 1Mb or 60 seconds.
Cost¶
- pay per data send
- near real time (60 sec minimum)
Destinations¶
- Destination Bucket - S3
- Any HTTP endpoints
- ElasticSearch
- Amazon Redshift
- OpenSearch
- Splunk
- Datadog
- New Relic
- Datadog
- MongoDb