We are currently running Elasticsearch 1.5.2 which we attempt to back up every midnight by issuing a
createSnapshot API call to a shared file system repo
backup. What puzzles us is that in our QA and production environment we only see the three or four latest snapshots being retained in the repository, while in our development environment and on local computers we tend to see a much higher number of snapshots being kept in the repo. We believe that QA and production also should show an ever increasing number of available snapshots and not be limited to three or four.
As far as I have been able to find out, Elasticsearch does not implement or offer any configuration of retention policy, number of snapshots to keep, auto delete or the like. I've searched through our code base and there are no usages of
deleteRepository that I can find.
Is there something regarding snapshot/restore we're missing or not understanding? Does anyone have an explanation for the behavior we're seeing?