Fluentd v0.12.8 is released

Hi users!

We have released Fluentd version 0.12.8. Here are the changes:

New features / Enhancement

  • output: Support millisecond flush span for try_flush_interval and queued_chunk_flush_interval: #568
  • filter_record_transformer: Support fields which start with @: #574
  • config: Add final attribute to prevent config overwritten by subclass: #583

Bug fixes

  • config: Fix embedded code parsing: #564
  • out_forward: Resend chunks on dropped connection when wait ACK: #580

Support millisecond flush span for try_flush_interval and queued_chunk_flush_interval

Buffers are now flushed within millisecond unit. In earlier versions, minimum buffer flush span was 1 second.

Default try_flush_interval and queued_chunk_flush_interval are still 1 second, but you can specify millisecond value like 0.1.
For example, DeNA uses try_flush_interval 0.1 and queued_chunk_flush_interval 0.01 parameters.

Note that this is useful for low latency data transfer but there is a trade-off between throughput and latency. Default values are enough on almost cases.



Happy logging!

There are some commercial supports for Fluentd, see Enterprise Services. If you use Fluentd on production, Let's share your use-case/testimonial on Testimonials page. Please consider to feedback Use-Case/Testimonials via GitHub.

Subscribed to the RSS feed here.

Written by Masahiro Nakagawa

Masahiro (@repeatedly) is the main maintainer of Fluentd. He works on Fluentd development and support full-time. He is also a committer of the D programming language.