## Updating custom templates to support `node_roles` and autoscalingedit

Custom deployment templates should be updated in order to take advantage of new Elastic Cloud Enterprise features, such as Data tiers (that is, the new cold and frozen data tiers) and Deployment autoscaling. By updating these templates we also ensure forward compatibility with future Elastic Cloud Enterprise versions that will require certain fields such as `node_roles` and `id` to be present in the deployment configuration.

System owned deployment templates have already been updated to support both data tiers with `node_roles` and autoscaling. However, the custom templates that you created need to be manually updated by following the steps in this guide.

The `node_roles` field defines the roles that an Elasticsearch topology element can have, which is used in place of `node_type` when a new feature such as autoscaling is enabled, or when a new data tier is added. This field is supported on Elastic stack versions 7.10 and above.

There are a number of fields that need to be added to each Elasticsearch node in order to support `node_roles`:

• id: Unique identifier of the topology element. This field, along with the `node_roles`, identifies an Elasticsearch topology element.
• node_roles: The list of node roles. Allowable roles are: `master`, `ingest`, `ml`, `data_hot`, `data_content`, `data_warm`, `data_cold`, `data_frozen`, `remote_cluster_client`, and `transform`. For details, see Node roles.
• topology_element_control: Controls for the topology element.

• min: The absolute minimum size limit for a topology element. If the value is `0`, that means the topology element can be disabled.

The following example is based on the `default` system owned deployment template that already supports `node_roles`. This template will be used as a reference for the next sections:

Reference example with support for `node_roles`
```{
...
"deployment_template": {
"resources": {
"elasticsearch": [
{
"plan": {
"cluster_topology": [
{
"id": "hot_content",
"instance_configuration_id": "data.default",
"zone_count": 1,
"node_roles": [
"master",
"ingest",
"data_hot",
"data_content",
"remote_cluster_client",
"transform"
],
"node_type": {
"master": true,
"data": true,
"ingest": true
},
"elasticsearch": {
"node_attributes": {
"data": "hot"
}
},
"size": {
"value": 4096,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 1024,
"resource": "memory"
}
}
},
{
"id": "warm",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_warm",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "warm"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "cold",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_cold",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "cold"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "frozen",
"instance_configuration_id": "data.frozen",
"zone_count": 1,
"node_roles": [
"data_frozen"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "frozen"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "coordinating",
"zone_count": 1,
"instance_configuration_id": "coordinating",
"node_roles": [
"ingest",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "master",
"zone_count": 1,
"instance_configuration_id": "master",
"node_roles": [
"master",
"remote_cluster_client"
],
"node_type": {
"master": true,
"data": false,
"ingest": false
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "ml",
"zone_count": 1,
"instance_configuration_id": "ml",
"node_roles": [
"ml",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": false,
"ml": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
}
],
"elasticsearch": {}
},
...
}
],
"kibana": [
{
"ref_id": "kibana-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"zone_count": 1,
"cluster_topology": [
{
"instance_configuration_id": "kibana",
"size": {
"value": 1024,
"resource": "memory"
}
}
],
"kibana": {}
}
}
],
"apm": [
{
"ref_id": "apm-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"instance_configuration_id": "apm",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 1
}
],
"apm": {}
}
}
],
"enterprise_search": [
{
"ref_id": "enterprise_search-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"node_type": {
"appserver": true,
"connector": true,
"worker": true
},
"instance_configuration_id": "enterprise.search",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 2
}
],
"enterprise_search": {}
}
}
]
}
}
}```

In the reference example there are seven different Elasticsearch topology elements: `hot_content`, `warm`, `cold`, `frozen`, `coordinating`, `master`, and `ml`. These names map to the `id` field of each topology element. In addition, this template contains four different resources: `elasticsearch`, `kibana`, `apm`, and `enterprise_search`.

##### Requirementsedit

To add support for `node_roles`, the deployment template has to meet the following requirements:

• Contains all four `resources`: `elasticsearch`, `kibana`, `apm`, and `enterprise_search`.
• The `elasticsearch` resource contains all seven topology elements: `hot_content`, `warm`, `cold`, `frozen`, `coordinating`, `master`, and `ml`.

The IDs `hot_content`, `warm`, `cold`, `frozen`, `coordinating`, `master`, and `ml` are the only ones supported in an Elasticsearch topology element. In addition, there may not be topology elements with duplicate IDs inside the `elasticsearch` resource.

• Each topology element contains the `id`, `node_roles`, and `topology_element_control` fields.

It is also recommended that all Elasticsearch topology elements have the `node_attributes` field. This field can be useful in ILM policies, when creating a deployment using a version below 7.10, that does not support `node_roles`.

Except for the `id` and `node_roles`, all fields can be configured by the user. Also, the topology elements must contain the exact same `id` and `node_roles` that are present in the reference example.

Although it is required for the template to contain all resources and topology elements, it is possible to disable certain components by setting their `size.value` (and `topology_element_control.size` in the case of the topology elements) to `0`.

##### Updating an ECE custom template to support `node_roles`edit

To update a custom deployment template:

1. Add the `id`, `node_roles`, `node_attributes`, and `topology_element_control` fields to the existing Elasticsearch topology elements. Keep in mind that these fields must match the Elasticsearch topology element in question:

• If the `id` of the topology elements in the existing templates already match any of the seven mentioned in the requirements, then simply add the `node_roles` and `topology_element_control` to those elements, based on the reference example.
• Otherwise, map each of the existing topology elements to one of the seven Elasticsearch topology elements, based on their `node_type`, and add the fields accordingly.
2. Add the `elasticsearch` topology elements that are missing.
3. Add the `resources` that are missing.
##### Exampleedit

The existing template contains three Elasticsearch topology elements and two resources (`elasticsearch` and `kibana`).

Custom example without support for `node_roles`
```{
...
"deployment_template": {
"resources": {
"elasticsearch": [
{
"plan": {
"cluster_topology": [
{
"instance_configuration_id": "custom.data",
"zone_count": 2,
"node_type": {
"master": true,
"data": true,
"ingest": true
},
"size": {
"value": 8192,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 1024,
"resource": "memory"
}
}
},
{
"zone_count": 1,
"instance_configuration_id": "custom.master",
"node_type": {
"master": true,
"data": false,
"ingest": false
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"zone_count": 1,
"instance_configuration_id": "custom.ml",
"node_type": {
"master": false,
"data": false,
"ingest": false,
"ml": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
}
],
"elasticsearch": {}
},
...
}
],
"kibana": [
{
"ref_id": "kibana-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"zone_count": 1,
"cluster_topology": [
{
"instance_configuration_id": "kibana",
"size": {
"value": 1024,
"resource": "memory"
}
}
],
"kibana": {}
}
}
]
}
}
}```

In this case we can match the three existing Elasticsearch topology elements to `hot_content`, `master`, and `ml`, respectively, based on their `node_type` field. Therefore, we can simply add the `id`, `node_roles`, `topology_element_control`, and `node_attributes` that are already associated with these topology elements in the reference example.

Then, it is only necessary to add the four Elasticsearch topology elements (`warm`, `cold`, `frozen`, and `coordinating`) and two resources (`apm` and `enterprise_search`) that are missing. These fields can also be added based on the reference example.

After adding support for `node_roles`, the resulting deployment template should look similar to the following:

Custom example with support for `node_roles`
```{
...
"deployment_template": {
"resources": {
"elasticsearch": [
{
"plan": {
"cluster_topology": [
{
"id": "hot_content",
"instance_configuration_id": "custom.data",
"zone_count": 2,
"node_roles": [
"master",
"ingest",
"data_hot",
"data_content",
"remote_cluster_client",
"transform"
],
"node_type": {
"master": true,
"data": true,
"ingest": true
},
"elasticsearch": {
"node_attributes": {
"data": "hot"
}
},
"size": {
"value": 8192,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 1024,
"resource": "memory"
}
}
},
{
"id": "warm",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_warm",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "warm"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "cold",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_cold",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "cold"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "frozen",
"instance_configuration_id": "data.frozen",
"zone_count": 1,
"node_roles": [
"data_frozen"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "frozen"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "coordinating",
"zone_count": 1,
"instance_configuration_id": "coordinating",
"node_roles": [
"ingest",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "master",
"zone_count": 1,
"instance_configuration_id": "custom.master",
"node_roles": [
"master",
"remote_cluster_client"
],
"node_type": {
"master": true,
"data": false,
"ingest": false
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "ml",
"zone_count": 1,
"instance_configuration_id": "custom.ml",
"node_roles": [
"ml",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": false,
"ml": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
}
],
"elasticsearch": {}
},
...
}
],
"kibana": [
{
"ref_id": "kibana-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"zone_count": 1,
"cluster_topology": [
{
"instance_configuration_id": "kibana",
"size": {
"value": 1024,
"resource": "memory"
}
}
],
"kibana": {}
}
}
],
"apm": [
{
"ref_id": "apm-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"instance_configuration_id": "apm",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 1
}
],
"apm": {}
}
}
],
"enterprise_search": [
{
"ref_id": "enterprise_search-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"node_type": {
"appserver": true,
"connector": true,
"worker": true
},
"instance_configuration_id": "enterprise.search",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 2
}
],
"enterprise_search": {}
}
}
]
}
}
}```

After adding support for `node_roles` we can then update the template to support autoscaling. Autoscaling is used to automatically adjust the available resources in the deployments. Currently, this feature is available for Elasticsearch data tiers and machine learning node in Elastic stack versions 7.11 and above.

There are a number of autoscaling fields that need to be added in order to support autoscaling:

• autoscaling_min: The default minimum size of an Elasticsearch topology element when autoscaling is enabled. This setting is currently available only for machine learning nodes, since these are the only nodes that support scaling down.
• autoscaling_max: The default maximum size of an Elasticsearch topology element when autoscaling is enabled. This setting is currently available only for data tiers and machine learning nodes, since these are the only nodes that support scaling up.
• autoscaling_enabled: When set to `true`, autoscaling is enabled by default on an Elasticsearch cluster.

These fields represent the default settings for the deployment. However, autoscaling can be enabled/disabled and the maximum and minimum autoscaling sizes can be adjusted in the deployment settings.

Similar to the `node_roles` example, the following one is also based on the `default` deployment template that already supports `node_roles` and autoscaling. This template will be used as a reference for the next sections:

Reference example with support for `node_roles` and autoscaling
```{
...
"deployment_template": {
"resources": {
"elasticsearch": [
{
"plan": {
"cluster_topology": [
{
"id": "hot_content",
"instance_configuration_id": "data.default",
"zone_count": 1,
"node_roles": [
"master",
"ingest",
"data_hot",
"data_content",
"remote_cluster_client",
"transform"
],
"node_type": {
"master": true,
"data": true,
"ingest": true
},
"elasticsearch": {
"node_attributes": {
"data": "hot"
}
},
"size": {
"value": 4096,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 1024,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "warm",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_warm",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "warm"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "cold",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_cold",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "cold"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "frozen",
"instance_configuration_id": "data.frozen",
"zone_count": 1,
"node_roles": [
"data_frozen"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "frozen"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "coordinating",
"zone_count": 1,
"instance_configuration_id": "coordinating",
"node_roles": [
"ingest",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "master",
"zone_count": 1,
"instance_configuration_id": "master",
"node_roles": [
"master",
"remote_cluster_client"
],
"node_type": {
"master": true,
"data": false,
"ingest": false
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "ml",
"zone_count": 1,
"instance_configuration_id": "ml",
"node_roles": [
"ml",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": false,
"ml": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_min": {
"resource": "memory",
"value": 0
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
}
],
"elasticsearch": {},
"autoscaling_enabled": true
},
...
}
],
"kibana": [
{
"ref_id": "kibana-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"zone_count": 1,
"cluster_topology": [
{
"instance_configuration_id": "kibana",
"size": {
"value": 1024,
"resource": "memory"
}
}
],
"kibana": {}
}
}
],
"apm": [
{
"ref_id": "apm-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"instance_configuration_id": "apm",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 1
}
],
"apm": {}
}
}
],
"enterprise_search": [
{
"ref_id": "enterprise_search-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"node_type": {
"appserver": true,
"connector": true,
"worker": true
},
"instance_configuration_id": "enterprise.search",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 2
}
],
"enterprise_search": {}
}
}
]
}
}
}```
##### Requirementsedit

To add support for autoscaling, the deployment template has to meet the following requirements:

1. Already has support for `node_roles`.
2. Contains the `size`, `autoscaling_min`, and `autoscaling_max` fields, according to the rules specified in the autoscaling requirements table.
3. Contains the `autoscaling_enabled` fields on the `elasticsearch` resource.

If necessary, the values chosen for each field can be based on the reference example.

##### Updating an ECE custom template to support autoscalingedit

To update a custom deployment template:

1. Add the `autoscaling_min` and `autoscaling_max` fields to the Elasticsearch topology elements (see Autoscaling through the API).
2. Add the `autoscaling_enabled` fields to the `elasticsearch` resource. Set this field to `true` in case you want autoscaling enabled by default, and to `false` otherwise.
##### Exampleedit

After adding support for autoscaling to the example presented in the previous section, the resulting deployment template should look similar to the following:

Custom example with support for `node_roles` and autoscaling
```{
...
"deployment_template": {
"resources": {
"elasticsearch": [
{
"plan": {
"cluster_topology": [
{
"id": "hot_content",
"instance_configuration_id": "custom.data",
"zone_count": 2,
"node_roles": [
"master",
"ingest",
"data_hot",
"data_content",
"remote_cluster_client",
"transform"
],
"node_type": {
"master": true,
"data": true,
"ingest": true
},
"elasticsearch": {
"node_attributes": {
"data": "hot"
}
},
"size": {
"value": 8192,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 1024,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "warm",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_warm",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "warm"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "cold",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_cold",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "cold"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "frozen",
"instance_configuration_id": "data.frozen",
"zone_count": 1,
"node_roles": [
"data_frozen"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "frozen"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "coordinating",
"zone_count": 1,
"instance_configuration_id": "coordinating",
"node_roles": [
"ingest",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "master",
"zone_count": 1,
"instance_configuration_id": "custom.master",
"node_roles": [
"master",
"remote_cluster_client"
],
"node_type": {
"master": true,
"data": false,
"ingest": false
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "ml",
"zone_count": 1,
"instance_configuration_id": "custom.ml",
"node_roles": [
"ml",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": false,
"ml": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_min": {
"resource": "memory",
"value": 0
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
}
],
"elasticsearch": {},
"autoscaling_enabled": true
},
...
}
],
"kibana": [
{
"ref_id": "kibana-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"zone_count": 1,
"cluster_topology": [
{
"instance_configuration_id": "kibana",
"size": {
"value": 1024,
"resource": "memory"
}
}
],
"kibana": {}
}
}
],
"apm": [
{
"ref_id": "apm-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"instance_configuration_id": "apm",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 1
}
],
"apm": {}
}
}
],
"enterprise_search": [
{
"ref_id": "enterprise_search-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"node_type": {
"appserver": true,
"connector": true,
"worker": true
},
"instance_configuration_id": "enterprise.search",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 2
}
],
"enterprise_search": {}
}
}
]
}
}
}```

#### Updating a custom template through the RESTful APIedit

Having added support for `node_roles` and autoscaling to your custom template, it is possible to perform the update through the RESTful API, by following these steps:

1. Obtain the existing deployment templates by sending the following `GET` request, and take note of the `id` of the template you wish to update.

`curl -k -X GET -H "Authorization: ApiKey $ECE_API_KEY" https://COORDINATOR_HOST:12443/api/v1/deployments/templates?region=ece-region` 2. Send a `PUT` request with the updated template on the payload, in order to effectively replace the outdated template with the new one. Note that the following request is just an example, you have to replace `{template_id}` with the `id` you collected on step 1. and set the payload to the updated template JSON. See set deployment template API for more details. Update template API request example ```curl -k -X PUT -H "Authorization: ApiKey$ECE_API_KEY" https://\$COORDINATOR_HOST:12443/api/v1/deployments/templates/{template_id}?region=ece-region -H 'content-type: application/json' -d '
{
"name": "ECE Custom Template",
"description": "ECE custom template with support for node_roles and autoscaling",
"deployment_template": {
"resources": {
"elasticsearch": [
{
"ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"id": "hot_content",
"instance_configuration_id": "custom.data",
"zone_count": 2,
"node_roles": [
"master",
"ingest",
"data_hot",
"data_content",
"remote_cluster_client",
"transform"
],
"node_type": {
"master": true,
"data": true,
"ingest": true
},
"elasticsearch": {
"node_attributes": {
"data": "hot"
}
},
"size": {
"value": 8192,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 1024,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "warm",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_warm",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "warm"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "cold",
"instance_configuration_id": "data.highstorage",
"zone_count": 1,
"node_roles": [
"data_cold",
"remote_cluster_client"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "cold"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "frozen",
"instance_configuration_id": "data.frozen",
"zone_count": 1,
"node_roles": [
"data_frozen"
],
"node_type": {
"data": true,
"ingest": false,
"master": false
},
"elasticsearch": {
"node_attributes": {
"data": "frozen"
}
},
"size": {
"resource": "memory",
"value": 0
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
},
{
"id": "coordinating",
"zone_count": 1,
"instance_configuration_id": "coordinating",
"node_roles": [
"ingest",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "master",
"zone_count": 1,
"instance_configuration_id": "custom.master",
"node_roles": [
"master",
"remote_cluster_client"
],
"node_type": {
"master": true,
"data": false,
"ingest": false
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
}
},
{
"id": "ml",
"zone_count": 1,
"instance_configuration_id": "custom.ml",
"node_roles": [
"ml",
"remote_cluster_client"
],
"node_type": {
"master": false,
"data": false,
"ingest": false,
"ml": true
},
"size": {
"value": 0,
"resource": "memory"
},
"topology_element_control": {
"min": {
"value": 0,
"resource": "memory"
}
},
"autoscaling_min": {
"resource": "memory",
"value": 0
},
"autoscaling_max": {
"value": 2097152,
"resource": "memory"
}
}
],
"elasticsearch": {},
"autoscaling_enabled": true
},
"settings": {
"dedicated_masters_threshold": 6
}
}
],
"kibana": [
{
"ref_id": "kibana-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"zone_count": 1,
"cluster_topology": [
{
"instance_configuration_id": "kibana",
"size": {
"value": 1024,
"resource": "memory"
}
}
],
"kibana": {}
}
}
],
"apm": [
{
"ref_id": "apm-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"instance_configuration_id": "apm",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 1
}
],
"apm": {}
}
}
],
"enterprise_search": [
{
"ref_id": "enterprise_search-ref-id",
"elasticsearch_cluster_ref_id": "es-ref-id",
"region": "ece-region",
"plan": {
"cluster_topology": [
{
"node_type": {
"appserver": true,
"worker": true,
"connector": true
},
"instance_configuration_id": "enterprise.search",
"size": {
"value": 0,
"resource": "memory"
},
"zone_count": 2
}
],
"enterprise_search": {}
}
}
]
}
},
"system_owned": false
}'```

After the template is updated, you can start creating new deployments or migrating existing ones to use the updated template.

Although `node_roles` and autoscaling are only available in more recent Elastic stack versions, an updated template can still be used with deployments that have versions below 7.10. In these cases, the data tiers and autoscaling features will only take effect once the deployment is upgraded to versions 7.10 and 7.11, respectively.

#### Migrating a deployment to the updated templateedit

Once a custom template is updated, the existing deployments associated with this template can be migrated to the updated template. This migration can occur in two different ways:

• By waiting for the next plan change to occur, allowing the migration to occur automatically.
• By applying a no-op plan, to force the migration to occur immediately.

A no-op plan can be applied by performing a rolling update on your deployment without applying any configuration changes. This can be done through the UI, with the following steps:

1. On the Deployments page, select your deployment.
2. From your deployment menu, go to Manage and select Edit deployment.
3. On the Edit page, click Save.

Once a deployment is migrated to the updated deployment template, it is not possible to roll back.

After the plan change has finished, it is recommended to follow the Migrate index allocation filters to node roles guide. Step 1 of this guide was already accomplished by adding support for `node_roles`. However, performing steps 2, 3, and 4, which involves updating index settings, index templates, and ILM policies, can prevent shard allocation issues caused by conflicting ILM policies.