LLM Assistant

Configure LLM features in Bindplane using Google Cloud Vertex AI

Bindplane's LLM features help streamline complex configuration tasks by providing intelligent suggestions and automated generation capabilities.

Note: LLM features are currently in Private Preview. Contact your administrator or support team for access.

Current Features

LLM-powered Regex Generation

The Parse with Regex processor includes a "Generate" button that helps create complex regex statements for telemetry parsing.

How does it work?

• Analyzes your live data to understand the parsing requirements

• Generates recommended regex patterns with capture groups

• Provides a strong starting point compared to writing regex from scratch

• Does not store or persist any of your data - it's only used for prompt generation

Privacy and Data Governance

• Bindplane does not store LLM requests or responses

• No context is saved between LLM requests and responses

• Your data is only used temporarily to generate prompts

• Review GCP Vertex AI data governance policy for additional details

Cloud Setup

In your Bindplane Cloud account, navigate to organization settings and switch on the LLM features toggle. This feature is not enabled by default and requires opt-in.

Self-Hosted Setup

Requirements

• Network connectivity to the Vertex AI API

• Google Cloud Platform project with Vertex AI API enabled

• GCP Service Account with appropriate permissions

Google Cloud Platform Configuration

1. Enable Vertex AI API in your GCP project

   • Follow the Cloud Environment Setup Guide

2. Create a Service Account with the Vertex AI User role

   • Find your Project ID

   • Create service account key file for authentication

   • Choose your location for generative AI resources

Bindplane Configuration

Add the following section to your Bindplane configuration file:

llm:
  gemini:
    projectID: my-project-id
    location: us-central1
    maxTokens: 1024 # Optional
    credentialsFile: /files/myCredentials.json # Optional

Configuration Parameters

Parameter	Required	Description
projectID	Yes	Your GCP Project ID
location	Yes	Location of generative AI resources (e.g., us-central1) Available Locations
maxTokens	No	Cap for generated tokens in responses (uses Gemini defaults if not specified)
credentialsFile	No	Absolute path to service account key file (uses Application Default Credentials if not specified)

For more information on authentication options, see Application Default Credentials.

Troubleshooting

• Network connectivity: Ensure your firewall allows outbound connections to the Vertex AI API endpoints

• Authentication: Verify your service account has the correct Vertex AI User role

• API access: Confirm the Vertex AI API is enabled in your GCP project

For additional support, contact your system administrator or Bindplane support.