CONFIG.md

Configuration

Configuration lives in config.yml. This file is created when you first run make install, and is "git ignored" to help prevent committing secrets. The default configs can all be seen in the config.yml.example, which is not intended to be modified.

Run the connector service on-prem

To run a self-managed connector client on your own infrastructure, you have two options:

• Run from source or Use Docker

You'll need to carry out the following steps:

1. Go to Kibana, Search > Create an Elasticsearch index. Select the service type of the connector you wish to use.

2. Create an API key to work with the connector. Use the Generate API key button under Configuration tab.

3. Configure your connector service application. You need to configure the following fields, and leave the rest as default.

   1. elasticsearch.host: Configure this to the Elasticsearch endpoint.
   2. elasticsearch.api_key: Configure the API key generated in step 2. If you run multiple connector clients, or customized connectors in one connector service, you can configure the API key of any connector here. This is recommended over utilizing elasticsearch.username/password so that access can be automatically and narrowly scoped.
   3. connectors: Configure per-connector configuration here. It should at least contain connector_id (You can find the connector_id in step 3 Deploy a connector under Configuration tab in Kibana) and service_type. You can also configure the api_key if the global elasticsearch configuration does not have access to the content index of this connector. One example:

   connectors:
     -
       connector_id: <connector_id>
       service_type: <service_type>
       api_key: <api_key>

4. Run the connector service application with

   make run

ℹ️ NOTE: API keys for connector clients

As of v8.10.0 you can configure multiple connectors in your config.yml file.

The Kibana UI enables you to create API keys that are scoped to a specific index/connector. If you don't create an API key for a specific connector, the top-level elasticsearch.api_key or elasticsearch.username:elasticsearch.password value is used.

If these top-level Elasticsearch credentials are not sufficiently privileged to write to individual connector indices, you'll need to create these additional, scoped API keys.

Example configuration

Use this example as a guide:

# Replace the values for `api_key`, `connector_id`, and `service_type` with the values you printed in Kibana.

elasticsearch:
  api_key: <key1>
  # Used to write data to .elastic-connectors and .elastic-connectors-sync-jobs
  # Any connectors without a specific `api_key` value will default to using this key
connectors:
  - connector_id: 1234
    service_type: <service_type>
    api_key: <key2>
    # Used to write data to the `search-*` index associated with connector 1234
    # You may have multiple connectors in your config file!
  - connector_id: 5678
    service_type: <service_type>
    api_key: <key3>
  # Used to write data to the `search-*` index associated with connector 5678
  - connector_id: abcd
    service_type: <service_type>
  # No explicit api key specified, so this connector will used <key1>

Connector service on Elastic Cloud

When you have an Enterprise Search deployment on Elastic Cloud (post 8.5.0), the connector service is automatically deployed. The connector service runs in native mode on Elastic Cloud, and the Elasticsearch connection configurations (i.e. host, username, password) are automatically configured. A special Cloud user (cloud-internal-enterprise_search-server) is used for the Elasticsearch connection. This user has proper privileges on the connector index (.elastic-connectors), the connector job index (.elastic-connectors-sync-jobs) and the connector content indices (search-*).

To run self-managed connector clients, you will need to run the connector service on-prem, as described in the first section.