Skip to content

This release adds a new S3 operator and fixes a bug within the fork operator.

Aug 28, 2025 · @raxyte · #5449

The from_s3 operator reads files from Amazon S3 with support for glob patterns, automatic format detection, and file monitoring.

from_s3 "s3://my-bucket/data/**.json"

The operator supports multiple authentication methods including default AWS credentials, explicit access keys, IAM role assumption, and anonymous access for public buckets:

from_s3 "s3://my-bucket/data.csv",
access_key=secret("AWS_ACCESS_KEY"),
secret_key=secret("AWS_SECRET_KEY")

For S3-compatible services, specify custom endpoints via URL parameters:

from_s3 "s3://my-bucket/data/**.json?endpoint_override=minio.example.com:9000&scheme=http"

Additional features include file watching for continuous ingestion, automatic file removal or renaming after processing, and path field injection to track source files in events.

Fix fork operator stopping after initial events

Section titled “Fix fork operator stopping after initial events”

Sep 1, 2025 · @raxyte · #5450

We fixed a bug where the fork operator would stop processing events after handling only the first few events, causing data loss in downstream pipeline stages.