[[managing-indices]]
== Managing Indices

*Index Management* enables you to view index settings,
mappings, and statistics and perform index-level operations.
These include refreshing, flushing, clearing the cache, force merging segments,
freezing indices, and more. *Index Management* provides a convenient way to
perform bulk operations on multiple indices.

To access this feature, go to *Management > {es} > Index Management*. 
If security is enabled,
you must have the `monitor` cluster privilege and the `view_index_metadata` 
and `manage` index privileges to view the data.  See 
{xpack-ref}/security-privileges.html[Security Privileges] for more
information.

*Index Management* uses badges to make users aware when an index is {ref}/frozen-indices.html[frozen], 
a {ref}/ccr-put-follow.html[follower index], 
or a {ref}/rollup-get-rollup-index-caps.html[rollup index]. 
Clicking a badge filters for all indices of that type. 

[role="screenshot"]
image::images/management_index_labels.png[Index Management UI]

Clicking the name of an index displays the index summary and provides access to
the index settings, mapping, and statistics. 

From the *Manage* menu, you can perform these index-level operations on either 
a single or multiple indices:

* *Close the index*. Blocks the index from read/write operations. 
A closed index exists in the cluster, but doesn't consume resources 
other than disk space. If you reopen a closed index, it goes through the 
normal recovery process. 

* *Force merge the index*. Reduces the number of segments in your shard by 
merging smaller files and clearing deleted ones. Only force merge a read-only index.

* *Refresh the index*. Writes the operations in the indexing buffer to the 
filesystem cache. This is automatically done once per second. Forcing a manual 
refresh is useful during testing, but should not be routinely done in 
production because it has a performance impact.

* *Clear the index cache*. Clears all caches associated with the index. 

* *Flush the index*. Frees memory by syncing the filesystem cache to disk and 
clearing the cache. Once the sync is complete, the internal transaction log is reset.

* *Freeze the index*. Makes the index read-only and reduces its memory footprint 
by moving shards to disk. {ref}/frozen-indices.html[Frozen indices] remain 
searchable, but queries take longer.

* *Delete the index*. Permanently removes the index and all of it's documents.

* *Add an lifecycle policy*.  Specifies a policy for managing the lifecycle of the 
index.

For information about the available management operations,
see {ref}/indices.html[Indices APIs].