Curator: Tending your time-series indices in Elasticsearch
Switch branches/tags
Nothing to show
Clone or download
Pull request Compare This branch is 1401 commits behind elastic:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Have time-series indices in Elasticsearch? This is the tool for you!


Install dependencies

pip install -r requirements.txt

See python --help for usage specifics.


The default values for host, port and prefix are:

--host localhost
--port 9200
-t (or --timeout) 30
-C (or --curation-style) time
-T (or --time-unit) days
-p (or --prefix) logstash-
-s (or --separator) .
--max_num_segments 2

If your values match these you do not need to include them. The prefix should be everything before the date string.


Close indices older than 14 days, delete indices older than 30 days (See

python --host my-elasticsearch -d 30 -c 14

Keep 14 days of logs in elasticsearch:

python --host my-elasticsearch -d 14

Disable bloom filter for indices older than 2 days, close indices older than 14 days, delete indices older than 30 days:

python --host my-elasticsearch -b 2 -c 14 -d 30

Optimize (Lucene forceMerge) indices older than 2 days to 1 segment per shard:

python --host my-elasticsearch -t 3600 -o 2 --max_num_segments 1

Keep 1TB of data in elasticsearch, show debug output:

python --host my-elasticsearch -C space -g 1024 -D

Dry run of above:

python --host my-elasticsearch -C space -g 1024 -D -n

Documentation and Errata

If you need to close and delete based on different criteria, please use separate command lines, e.g.

python --host my-elasticsearch -C space -g 1024
python --host my-elasticsearch -c 15

When using optimize the current behavior is to wait until the optimize operation is complete before continuing. With large indices, this can result in timeouts with the default 30 seconds. It is recommended that you increase the timeout to at least 3600 seconds, if not more.


  • fork the repo
  • make changes in your fork
  • send a pull request!