kibana/docs/concepts/index-patterns.asciidoc

[[index-patterns]]
=== Create an index pattern

{kib} requires an index pattern to access the {es} data that you want to explore.
An index pattern selects the data to use and allows you to define properties of the fields.

An index pattern can point to a specific index, for example, your log data from yesterday,
or all indices that contain your data.  It can also point to a
{ref}/data-streams.html[data stream] or {ref}/indices-aliases.html[index alias].

You’ll learn how to:

* Create index patterns
* Set the default index pattern
* Delete index patterns

[float]
[[index-patterns-read-only-access]]
=== Before you begin

* To access the *Index Patterns* view, you must have the {kib} privilege
`Index Pattern Management`.  To create an index pattern, you must have the {es} privilege
`view_index_metadata`. To add the privileges, open the main menu, then click *Stack Management > Roles*.

* If a read-only indicator appears in {kib}, you have insufficient privileges
to create or save index patterns. The buttons to create new index patterns or
save existing index patterns are not visible. For more information,
refer to <<xpack-security-authorization,Granting access to {kib}>>.

[float]
[[settings-create-pattern]]
=== Create an index pattern

If you collected data using one of the {kib} <<connect-to-elasticsearch,ingest options>>, uploaded a file, or added sample data,
you get an index pattern for free, and can start exploring your data.
If you loaded your own data, follow these steps to create an index pattern.

. Open the main menu, then click to *Stack Management > Index Patterns*.

. Click *Create index pattern*.
+
[role="screenshot"]
image:management/index-patterns/images/create-index-pattern.png["Create index pattern"]

. Start typing in the *Index pattern* field, and {kib} looks for the names of
{es} indices that match your input.
** Use a wildcard (*) to match multiple indices.
For example, suppose your system creates indices for Apache data
using the naming scheme `filebeat-apache-a`, `filebeat-apache-b`, and so on.
An index pattern named `filebeat-a` matches a single source, and `filebeat-*` matches multiple data sources.
Using a wildcard is the most popular approach.

** Select multiple indices by entering multiple strings,
separated with a comma. Make sure there is no space after the comma.
For example, `filebeat-a,filebeat-b` matches two indices, but not other indices
you might have afterwards (filebeat-c).

** Use a minus sign (-) to exclude an index, for example, test*,-test3.

. Click *Next step*.

. If {kib} detects an index with a timestamp, expand the *Time field* menu,
and then specify the default field for filtering your data by time.
+
If your index doesn’t have time-based data, or if you don’t want to select
the default timestamp field, choose *I don’t want to use the Time Filter*.
+
NOTE: If you don’t set a default time field, you will not be able to use
global time filters on your dashboards. This is useful if
you have multiple time fields and want to create dashboards that combine visualizations
based on different timestamps.

. Click *Create index pattern*.
+
{kib} is now configured to use your {es} data.

. Select this index pattern when you search and visualize your data.

[float]
[[rollup-index-pattern]]
==== Create an index pattern for rolled up data

An index pattern can match one rollup index.  For a combination rollup
index pattern with both raw and rolled up data, use the standard notation:

```ts
rollup_logstash,kibana_sample_data_logs
```
For an example, refer to <<rollup-data-tutorial,Create and visualize rolled up data>>.

[float]
[[management-cross-cluster-search]]
==== Create an index pattern that searches across clusters

If your {es} clusters are configured for {ref}/modules-cross-cluster-search.html[{ccs}],
you can create an index pattern to search across the clusters of your choosing. Use the
same syntax that you'd use in a raw {ccs} request in {es}:

```ts
<cluster-names>:<pattern>
```

For example, to query {ls} indices across two {es} clusters
that you set up for {ccs}, named `cluster_one` and `cluster_two`,
use this for your index pattern:

```ts
 cluster_one:logstash-*,cluster_two:logstash-*
```

You can use wildcards in your cluster names
to match any number of clusters.  For example, to search {ls} indices across
clusters named `cluster_foo`, `cluster_bar`, and so on, create this index pattern:

```ts
cluster_*:logstash-*
```

To query across all {es} clusters that have been configured for {ccs},
use a standalone wildcard for your cluster name in your index
pattern:

```ts
*:logstash-*
```

You can use exclusions to exclude indices that might contain mapping errors.
To match indices starting with `logstash-`, and exclude those starting with `logstash-old` from
all clusters having a name starting with `cluster_`, you can use `cluster_*:logstash-*,cluster*:logstash-old*`.
To exclude a cluster, use `cluster_*:logstash-*,cluster_one:-*`.

Once an index pattern is configured using the {ccs} syntax, all searches and
aggregations using that index pattern in {kib} take advantage of {ccs}.

[float]
[[delete-index-pattern]]
=== Delete index patterns

When you delete an index pattern, you are unable to recover the associated field formatters, scripted fields, source filters,
and field popularity data. Deleting an index pattern does not remove any indices or data documents from {es}.

WARNING: Deleting an index pattern breaks all visualizations, saved searches, and other saved objects that reference the index pattern.

. Open the main menu, then click *Stack Management > Index Patterns*.

. Click the index pattern you want to delete.

. Delete (image:management/index-patterns/images/delete.png[Delete icon]) the index pattern.

[float]
[[reload-fields]]
=== What’s next

Learn how to <<managing-index-patterns,manage the data fields>> in your index patterns.