Getting Started with Netra

Welcome to Netra! This guide will walk you through the essential steps to connect your ClickHouse logs to the Netra platform securely. By following these instructions, you'll be able to leverage Netra's powerful analytics capabilities on your LLM-generated data while ensuring the privacy and security of your users' information.

How It Works

Netra provides insights, configuration recommendations, and code changes to optimize your AI-Native (LLM, Token-Based) applications by analyzing your existing logs. The process is designed to be secure and straightforward:

Ingest Data

1. Create a Secure Integration: You'll start by creating a dedicated, read-only user for Netra in your ClickHouse DB or using Bulk Ingest. This ensures that our platform can only access the specific data you authorize.
2. Optionally Clean Your Data: Remove Personally Identifiable Information (PII) from the logs before Netra ingests them. Examples of how to do this include:
   • SQL-Based Cleaning: A native ClickHouse solution that uses a Materialized View to automatically clean and forward logs to a new, secure table. Clean Log Table Guide
   • Python-Based Cleaning: A high-accuracy script that uses Microsoft's Presidio library for advanced PII detection and redaction, suitable for more complex cleaning requirements. Python PII Cleaning Script Guide
3. Grant Access: Once your data is cleaned and ready, you'll grant the Netra user access to the sanitized table.

MCP Integration

1. Add MCP Server: Add the MCP server to your IDE's MCP settings file. Add MCP Server

Activate

1. System-Wide Tuning System-Wide Tuning
2. Real-time Middleware Real-time Middleware

Next Steps

To begin, please follow our step-by-step guides in the recommended order:

1. Create a Secure ClickHouse User for Netra Ingest: Your first step is to set up the necessary read-only permissions for the integration.
2. Choose a Cleaning Method:
   • For a quick, SQL-native approach, follow the Clean Log Table Guide.
   • For the most comprehensive PII redaction, use the Python PII Cleaning Script Guide.

After completing these steps, you will have a secure, clean, and continuous pipeline of your LLM logs ready for analysis in Netra.