Be notified of new releases
Create your free GitHub account today to subscribe to this repository for new releases and build software alongside 28 million developers.Sign up
Per this StackOverlow answer, it's probably a better idea to use
index for _bulk API calls than
create. Create is for strict creation and may fail if the doc already exists.
index is more of an upsert and ES will decide if the document should be created or updated. It's a better guarantee the call will succeed.
This release adds support (finally) for the Elasticsearch _bulk API. When setting the BufferSize in the appender setup, if you specifiy a BufferSize > 1, the appender will automatically use the bulk API when flushing the buffer instead of individual index/create calls. This will result in much fewer calls to Elasticsearch and a more efficient setup for your application. See updated documentation accordingly.
Potentially breaking changes in this release
- Dropped all dependencies on support libraries like NEST and JSON.Net. These were initially helpful but they are overkill for what is needed for the lib.
- Changed the TimeStamp field to use UTC time format to make it friendlier to use.
- Modified the rolling index name format to use YYYY.MM.DD instead of YYYY-MM-DD to be more logstash-ey
- Added "@timestamp" field to the Properties array with UTC format
Thanks to @kjersti for the date field suggestions!
Nuget package has been updated appropriately.