Amazon Data Firehose
- What it is
- Fully managed service that captures streaming data and delivers it to a destination — no code, no shards, no scaling to manage.
- Formerly Kinesis Data Firehose.
- Destinations
- Amazon S3, Amazon Redshift, Amazon OpenSearch Service, Splunk, and generic HTTP endpoints.
- What it does on the way
- Buffers (by size or time, typically ~60 seconds), compresses, encrypts, and can transform records with a AWS Lambda function.
- Can convert JSON to Parquet or ORC for cheaper, faster querying with Amazon Athena.
- The key trade-off vs Amazon Kinesis Data Streams
- Firehose = near real time, zero management, no replay, one destination.
- Data Streams = true real time (milliseconds), you write the consumer, retention and replay, many consumers.
- Exam signal
- “Simplest way to load streaming data into S3/Redshift” → Firehose.