Vector v0.39.0 release notes

The Vector team is pleased to announce version 0.39.0!

Be sure to check out the upgrade guide for breaking changes in this release.

This release just contains a mix of small enhancements and bug fixes. See the changelog below.

Upgrading Vector
When upgrading, we recommend stepping through minor versions as these can each contain breaking changes while Vector is pre-1.0. These breaking changes are noted in their respective upgrade guides.



8 enhancements

  • VRL was updated to v0.16.0. This includes the following changes:

    Breaking Changes & Upgrade Guide:

    New Features:


  • The kafka sink has a new healthcheck_topic configuration option to configure the topic used for healthchecks. Previously, topic was used to perform healthchecks, but a templated value would cause healthchecks to fail. Thanks to yalinglee for contributing this change!
  • The namespace field for the gcp_chronicle sink is now a templatable field. Thanks to chocpanda for contributing this change!
  • The prometheus_exporter sink now supports compressing responses based on the `Accept-Encoding header sent with client requests. Thanks to edhjer for contributing this change!
  • A pretty option to the json codec to output events as prettified JSON. Thanks to lsampras for contributing this change!
  • The http_server source can now optionally annotate events with the remote IP by setting the host_key option. Thanks to NishantJoshi00 for contributing this change!
  • Vector’s start-up time was greatly improved when loading configurations including many remap transforms. Thanks to Zettroke for contributing this change!
  • The Alpine base image used for Vector -alpine and -distroless-static images was updated to 3.20.

3 new features

  • The kafka source will now consume less memory by not over allocating buffers while reading. Thanks to biuboombiuboom for contributing this change!
  • The opentelemetry source now has experimental support for ingesting traces. Thanks to caibirdme for contributing this change!
  • A new configuration option end_every_period_ms is available on reduce transforms If supplied, every time this interval elapses for a given grouping, the reduced value for that grouping is flushed. Checked every flush_period_ms. Thanks to charlesconnell for contributing this change!

2 bug fixes

  • The endpoint in the Datadog sinks is rewritten the same as the Datadog Agent does to produce the API validation endpoint to avoid a HTTP 404 Not Found error when running the health check.
  • The kafka source main loop has been biased to handle acknowledgements before new messages to avoid a memory leak under high load.

Download Version 0.39.0

Linux (deb)
Linux (rpm)
Windows (MSI)