2014年09月25日 エンジニアリング

Elasticsearch Curator Version 2.0 Released

By Aaron Mildenstein

I am pleased to announce the immediate availability of Curator 2.0! Curator is our tool that helps your curate, or manage, your time-series indices.

What's changed?

So many new features were being added to Curator that it's monolithic nature was getting in the way of its future progress. As a result, Curator 2.0 now separates the Elasticsearch Curator Python API from the Curator Script.

Because there have been some changes in the command-line flags, be sure to test with the --dry-run flag.

The Elasticsearch Curator Python API

As the API has been separated, it is now simple to write your own scripts using the same methods that Curator uses. These methods are documented thoroughly here.

The Curator Script

In order to preserve reverse compatibility as much as possible, installing Curator in the recommended way (pip install elasticsearch-curator) continues to install an entry_point that allows Curator to be invoked with the same command.

curator delete --older-than 30 will work exactly as it did before!

New Features

The changes to the Elasticsearch Curator Python API aren't the only new features!

Snapshots

Snapshot functionality has been completely reworked to allow multiple indices per snapshot, incremental snapshots, named snapshots, and the ability to capture _all indices in a snapshot. Accordingly, when you wish to delete snapshots older than a given time period, Curator will now use the time the snapshot was created (as stored in the snapshot metadata) to determine its age.

Prefixes and suffixes and wildcards…Oh, my!

In addition to prefixes, Curator now allows you to use suffixes and wildcards to determine the index pattern it should look for. Additionally, prefixes and suffixes can be empty, allowing your index name to be just a date. Curator now expects your indices to match a pattern of prefix + timestring + suffix. Both prefix and suffix support wildcards, and can be empty.

Delay after optimize

I love getting feature requests and bug reports from the community. One such feature request was to introduce a delay after an optimize. This delay allows your Elasticsearch cluster to quiesce before continuing with the next optimize call.

Conclusion

The new changes in Curator are awesome! I look forward to hearing about new ways you are using the Elasticsearch Curator API. If you run into trouble or find something we missed, please log an issue on our GitHub Issues page. If you love Curator, please tell us about it! We love tweets with #elasticsearch in them!

Curator is growing and maturing into a bigger, better project with each new release! Thanks for reading, and Happy Curating!