Skip to content
NeuBird AI Docs NeuBird AI Docs NeuBird AI Docs

Elasticsearch

Connecting NeuBird to your Elasticsearch instance enables collection of telemetry data related to search queries, index performance, and cluster health. To integrate, you authenticate using an Elasticsearch API key with read-only privileges.

Step 1: Generate an Elasticsearch API Key

Section titled “Step 1: Generate an Elasticsearch API Key”

  1. Access Elasticsearch Security Settings: Log in to your Elasticsearch instance, navigate to your deployment, and click on Open Kibana.

    Elastic Cloud deployment page with Open Kibana button

    On Kibana, navigate to the Management section.

    Kibana sidebar with Management section highlighted

  2. Navigate to API Keys: Under the Security settings, click on API Keys > Create API key.

    Security settings page with Create API key button

    Tip: If you can’t find the API Keys section from the Kibana sidebar, use the search bar at the top of the page and search for “api key” to jump directly to Stack Management > API keys.

    Stack Management API keys page, reached by searching "api key" in the Kibana top bar

  3. Create a New API Key:

    • Provide a name for the API key.

      API key creation form with name field

    • Select the type of API key and choose read-only as the security privilege level.

  4. Copy the API Key: Once generated, copy the API key. You will not be able to view it again after navigating away, so store it securely.

    Generated API key displayed with copy option

Connection Fields

Section titled “Connection Fields”
FieldRequiredDescription
NameYesA descriptive name for the connection
DescriptionNoA brief description of this integration
API KeyYesThe read-only API key generated in Step 1
Unique ID KeyNoOrganization-specific ID key for tracking and audits
Elasticsearch URLYesYour Elasticsearch instance URL (format: https://<your-elasticsearch-domain>/_api/<endpoint>)
Cloud IDNoThe Cloud ID if using Elastic Cloud
IndicesNoComma-separated index names to check when validating the connection, e.g. logs-*,metrics-*. Each entry is matched as a prefix, not an exact match — it’s treated as <value>* unless you already added the trailing * yourself, and both the plain index name and its data-stream backing-index form (.ds-<value>*) are checked. This only confirms NeuBird can reach at least one matching index when the connection is created; it does not restrict which indices the AI can query afterward — searches, mappings, and index listings can still see every index your API key has permission to read. Leave blank to validate against all open indices in the cluster instead.

For general instructions on creating and managing connections, see the Connections Overview.