Connect to Google Vertex

This page provides step-by-step instructions for connecting Elastic Security to Vertex AI’s large language models (LLMs). You’ll first need to enable Vertex AI, then generate a key, and finally connect your Elastic Security project to Vertex AI.

Important

Before continuing, you should have an active project in one of Google Vertex AI’s supported regions.

Enable the Vertex AI API

  1. Log in to the GCP console and navigate to Vertex AI → Vertex AI Studio → Overview.
  2. If you’re new to Vertex AI, the Get started with Vertex AI Studio popup appears. Click Vertex AI API, then click ENABLE.

The following video demonstrates these steps.

connect-vertex-api-video

Note

For more information about enabling the Vertex AI API, refer to Google’s documentation.

Create a Vertex AI service account

  1. In the GCP console, navigate to APIs & Services → Library.
  2. Search for Vertex AI API, select it, and click MANAGE.
  3. In the left menu, navigate to Credentials then click + CREATE CREDENTIALS and select Service account.
  4. Name the new service account, then click CREATE AND CONTINUE.
  5. Under Select a role, select Vertex AI User, then click CONTINUE.
  6. Click Done.

The following video demonstrates these steps.

create-vertex-account-video

Generate a key

  1. Return to Vertex AI’s Credentials menu and click Manage service accounts.
  2. Search for the service account you just created, select it, then click the link that appears under Email.
  3. Go to the KEYS tab, click ADD KEY, then select Create new key.
  4. Select JSON, then click CREATE to download the key. Keep it somewhere secure.

The following video demonstrates these steps.

create-vertex-key-video

Connect Kibana to Google Vertex AI

Create an inference endpoint so that Kibana can use models from Google Vertex AI. AI features in Kibana can use only endpoints with the chat_completion task type. For more information about adding endpoints, refer to External inference.

  1. Go to the External inference page by using the navigation menu or the global search field.
  2. Select Add endpoint.
  3. Select Google Vertex AI as the service.
  1. For Model ID, enter a Gemini model ID such as gemini-1.5-flash. Learn more about the models.
  2. For GCP Region and GCP Project, enter your region and project ID.
  3. For Credentials JSON, enter the key JSON that you generated.
  4. Select the chat_completion task type, then select Save.

Create a Google Gemini connector so that Kibana can use Gemini models on Vertex AI.

  1. Log in to your Elastic deployment.
  2. Find the Connectors page in the navigation menu or use the global search field. Then click Create Connector, select Google Gemini.
  3. Name your connector to help keep track of the model version you are using.
  4. Under URL, enter the URL for your region.
  5. Enter your GCP Region and GCP Project ID.
  6. Under Default model, specify either gemini-1.5.pro or gemini-1.5-flash. Learn more about the models.
  7. Under Authentication, enter your credentials JSON.
  8. Click Save.

The following video demonstrates these steps.

configure-gemini-connector-video