Create or update index template APIedit

Creates or updates an index template. Index templates define settings, mappings, and aliases that can be applied automatically to new indices.

response = client.indices.put_index_template(
  name: 'template_1',
  body: {
    index_patterns: [
      'te*'
    ],
    priority: 1,
    template: {
      settings: {
        number_of_shards: 2
      }
    }
  }
)
puts response
PUT /_index_template/template_1
{
  "index_patterns" : ["te*"],
  "priority" : 1,
  "template": {
    "settings" : {
      "number_of_shards" : 2
    }
  }
}

Requestedit

PUT /_index_template/<index-template>

Prerequisitesedit

  • If the Elasticsearch security features are enabled, you must have the manage_index_templates or manage cluster privilege to use this API.

Descriptionedit

Elasticsearch applies templates to new indices based on an wildcard pattern that matches the index name.

Index templates are applied during data stream or index creation. For data streams, these settings and mappings are applied when the stream’s backing indices are created.

Settings and mappings specified in a create index request override any settings or mappings specified in an index template.

Changes to index templates do not affect existing indices, including the existing backing indices of a data stream.

Comments in index templatesedit

You can use C-style /* */ block comments in index templates. You can include comments anywhere in the request body, except before the opening curly bracket.

Path parametersedit

<index-template>
(Required, string) Name of the index template to create.

Query parametersedit

create
(Optional, Boolean) If true, this request cannot replace or update existing index templates. Defaults to false.
master_timeout
(Optional, time units) Period to wait for a connection to the master node. If no response is received before the timeout expires, the request fails and returns an error. Defaults to 30s.

Request bodyedit

composed_of
(Optional, array of strings) An ordered list of component template names. Component templates are merged in the order specified, meaning that the last component template specified has the highest precedence. See Composing multiple component templates for an example.
data_stream

(Optional, object) If this object is included, the template is used to create data streams and their backing indices. Supports an empty object.

Data streams require a matching index template with a data_stream object. See create an index template.

Properties of data_stream
allow_custom_routing
(Optional, Boolean) If true, the data stream supports custom routing. Defaults to false.
hidden
(Optional, Boolean) If true, the data stream is hidden. Defaults to false.
index_mode

(Optional, string) Type of data stream to create. Valid values are null (regular data stream) and time_series (time series data stream).

If time_series, each backing index has an index.mode index setting of time_series.

index_patterns

(Required, array of strings) Array of wildcard (*) expressions used to match the names of data streams and indices during creation.

Elasticsearch includes several built-in index templates. To avoid naming collisions with these templates, see Avoid index pattern collisions.

_meta
(Optional, object) Optional user metadata about the index template. May have any contents. This map is not automatically generated by Elasticsearch.
priority
(Optional, integer) Priority to determine index template precedence when a new data stream or index is created. The index template with the highest priority is chosen. If no priority is specified the template is treated as though it is of priority 0 (lowest priority). This number is not automatically generated by Elasticsearch.
template

(Optional, object) Template to be applied. It may optionally include an aliases, mappings, or settings configuration.

Properties of template
aliases

(Optional, object of objects) Aliases to add.

If the index template includes a data_stream object, these are data stream aliases. Otherwise, these are index aliases. Data stream aliases ignore the index_routing, routing, and search_routing options.

Properties of aliases objects
<alias>

(Required, object) The key is the alias name. Index alias names support date math.

The object body contains options for the alias. Supports an empty object.

Properties of <alias>
filter
(Optional, Query DSL object) Query used to limit documents the alias can access.
index_routing
(Optional, string) Value used to route indexing operations to a specific shard. If specified, this overwrites the routing value for indexing operations.
is_hidden
(Optional, Boolean) If true, the alias is hidden. Defaults to false. All indices for the alias must have the same is_hidden value.
is_write_index
(Optional, Boolean) If true, the index is the write index for the alias. Defaults to false.
routing
(Optional, string) Value used to route indexing and search operations to a specific shard.
search_routing
(Optional, string) Value used to route search operations to a specific shard. If specified, this overwrites the routing value for search operations.
mappings

(Optional, mapping object) Mapping for fields in the index. If specified, this mapping can include:

See Mapping.

settings
(Optional, index setting object) Configuration options for the index. See Index Settings.
version
(Optional, integer) Version number used to manage index templates externally. This number is not automatically generated by Elasticsearch.
deprecated
(Optional, boolean) Marks this index template as deprecated. When creating or updating a non-deprecated index template that uses deprecated components, Elasticsearch will emit a deprecation warning.

Examplesedit

Index template with index aliasesedit

You can include index aliases in an index template.

PUT _index_template/template_1
{
  "index_patterns" : ["te*"],
  "template": {
    "settings" : {
        "number_of_shards" : 1
    },
    "aliases" : {
        "alias1" : {},
        "alias2" : {
            "filter" : {
                "term" : {"user.id" : "kimchy" }
            },
            "routing" : "shard-1"
        },
        "{index}-alias" : {} 
    }
  }
}

the {index} placeholder in the alias name will be replaced with the actual index name that the template gets applied to, during index creation.

Multiple matching templatesedit

If multiple index templates match the name of a new index or data stream, the template with the highest priority is used. For example:

PUT /_index_template/template_1
{
  "index_patterns" : ["t*"],
  "priority" : 0,
  "template": {
    "settings" : {
      "number_of_shards" : 1,
      "number_of_replicas": 0
    },
    "mappings" : {
      "_source" : { "enabled" : false }
    }
  }
}

PUT /_index_template/template_2
{
  "index_patterns" : ["te*"],
  "priority" : 1,
  "template": {
    "settings" : {
      "number_of_shards" : 2
    },
    "mappings" : {
      "_source" : { "enabled" : true }
    }
  }
}

For indices that start with te*, _source will enabled, and the index will have two primary shards and one replica, because only template_2 will be applied.

Multiple templates with overlapping index patterns at the same priority are not allowed, and an error will be thrown when attempting to create a template matching an existing index template at identical priorities.

Template versioningedit

You can use the version parameter to add a version number to an index template. External systems can use these version numbers to simplify template management.

The version parameter is optional and not automatically generated or used by Elasticsearch.

To unset a version, replace the template without specifying one.

response = client.indices.put_index_template(
  name: 'template_1',
  body: {
    index_patterns: [
      'foo',
      'bar'
    ],
    priority: 0,
    template: {
      settings: {
        number_of_shards: 1
      }
    },
    version: 123
  }
)
puts response
PUT /_index_template/template_1
{
  "index_patterns" : ["foo", "bar"],
  "priority" : 0,
  "template": {
    "settings" : {
        "number_of_shards" : 1
    }
  },
  "version": 123
}

To check the version, you can use the get index template API.

Template metadataedit

You can use the _meta parameter to add arbitrary metadata to an index template. This user-defined object is stored in the cluster state, so keeping it short is preferable.

The _meta parameter is optional and not automatically generated or used by Elasticsearch.

To unset _meta, replace the template without specifying one.

response = client.indices.put_index_template(
  name: 'template_1',
  body: {
    index_patterns: [
      'foo',
      'bar'
    ],
    template: {
      settings: {
        number_of_shards: 3
      }
    },
    _meta: {
      description: 'set number of shards to three',
      serialization: {
        class: 'MyIndexTemplate',
        id: 17
      }
    }
  }
)
puts response
PUT /_index_template/template_1
{
  "index_patterns": ["foo", "bar"],
  "template": {
    "settings" : {
        "number_of_shards" : 3
    }
  },
  "_meta": {
    "description": "set number of shards to three",
    "serialization": {
      "class": "MyIndexTemplate",
      "id": 17
    }
  }
}

To check the _meta, you can use the get index template API.

Data stream definitionedit

To use an index template for data streams, the template must include a data_stream object. See create an index template.

response = client.indices.put_index_template(
  name: 'template_1',
  body: {
    index_patterns: [
      'logs-*'
    ],
    data_stream: {}
  }
)
puts response
PUT /_index_template/template_1
{
  "index_patterns": ["logs-*"],
  "data_stream": { }
}

Composing aliases, mappings, and settingsedit

When multiple component templates are specified in the composed_of field for an index template, they are merged in the order specified, meaning that later component templates override earlier component templates. Any mappings, settings, or aliases from the parent index template are merged in next. Finally, any configuration on the index request itself is merged.

In this example, the order of the two component templates changes the number of shards for an index:

PUT /_component_template/template_with_2_shards
{
  "template": {
    "settings": {
      "index.number_of_shards": 2
    }
  }
}

PUT /_component_template/template_with_3_shards
{
  "template": {
    "settings": {
      "index.number_of_shards": 3
    }
  }
}

PUT /_index_template/template_1
{
  "index_patterns": ["t*"],
  "composed_of": ["template_with_2_shards", "template_with_3_shards"]
}

In this case, an index matching t* will have three primary shards. If the order of composed templates were reversed, the index would have two primary shards.

Mapping definitions are merged recursively, which means that later mapping components can introduce new field mappings and update the mapping configuration. If a field mapping is already contained in an earlier component, its definition will be completely overwritten by the later one.

This recursive merging strategy applies not only to field mappings, but also root options like dynamic_templates and meta. If an earlier component contains a dynamic_templates block, then by default new dynamic_templates entries are appended onto the end. If an entry already exists with the same key, then it is overwritten by the new definition.