Skip to content

Latest commit

 

History

History
32 lines (24 loc) · 3.93 KB

data-prepper.release-notes-1.1.0.md

File metadata and controls

32 lines (24 loc) · 3.93 KB

2021-xx-xx Version 1.1.0


Note: These commits were originally contributed to the Open Distro for Elasticsearch Data Prepper repo, they were migrated to this repo in #191. This set of release notes links back to the original ODFE pull requests for posterity.

Features

  • AWS Certificate Manager integration for both the OTel Trace Source and Peer Forwader plugins. Allows private keys and certificates to be fetched from ACM when establishing connections using TLS. (#705, #726)
  • AWS S3 integration for both the OTel Trace Source and Peer Forwader plugins. Allows private keys and certificates to be fetched from S3 when establishing connections using TLS. (#711, #726)
  • AWS Cloud Map integration. Cloud Map can be used as a peer discovery provider when operating a Data Prepper cluster. (#734)

Enhancements

  • The PeerForwarder target port is now configurable. (#742)
  • The AWS IAM role to be assumed by the OpenSearch sink is now configurable. (#736)
  • Multiple Micrometer metric registries can now be configured. The default remains Prometheus, however an AWS CloudWatch registry provider has been added. (#585)
  • The OTel Trace Source's gRPC server can be configured to accept 'application/json' requests in addition to gRPC requests. (#607)

Bug Fixes

  • Increased number of threads that write to sinks (previously hardcoded to 1). (#681)
  • Increased hash ring vNode count to better distribute requests made by the Peer Forwarder plugin. (#806)
  • Fixed the OTel Trace Source's health check service to actually be configurable, previously it was always enabled regardless of config settings. (#733)
  • Provided a better exception message when an OpenSearch index alias name is already taken by a concrete index. (#588)

Refactoring

  • Created a dedicated threadpool to enable async peer requests. Previously requests were synchronous and performed poorly in clusters with high peer counts. (#777)
  • Changed MapDB usage from disk to memory to improve performance. A 10x speed improvement was observed. (#671)
  • Fixed a race condition in the Service Map Stateful Prepper which could cause threads to stop responding. (#671)
  • Improved GC profile, re-used Java objects to limit the number of survivor objects in JVM heap. (#671)
  • Added an index template version check on startup to ensure old versions of Data Prepper won't overwrite newer index templates. (#608)
  • Misc refactoring of OTel Trace Raw Prepper, removed the DelayQueue in favor of a cache to improve performance. (#681)