Ragzero: Distributed Document Search Service

Ragzero is a high-performance, multi-tenant document search service built with Go, Elasticsearch, Redis, and PostgreSQL. It is designed to handle millions of documents with sub-500ms latency.

Architecture Highlights

• Multi-Tenancy: Strict isolation using X-Tenant-Id and X-API-Key header enforcement.
• Eventual Consistency: Asynchronous indexing via internal worker pools for high write availability.
• Observability: Distributed tracing with OpenTelemetry and structured JSON logging.
• Caching: L2 query results caching using Redis (Cache-Aside pattern).
• Rate Limiting: Per-tenant rate limiting to prevent noisy neighbor issues.
• Search Engine: Elasticsearch with BM25 relevance ranking, fuzzy matching, and result highlighting.

Prerequisites

• Docker & Docker Compose
• Go 1.26+ (managed via asdf)

Getting Started

1. Start the services

make up

2. Architecture & Design

The complete system design, production readiness analysis, and technical deep-dives are available here:

• Architecture & Design Document

3. API Testing

Sample API requests and lifecycle tests are available in the docs/http/ directory. These can be used to verify system functionality, including fuzzy search and multi-tenancy isolation.

License & Copyright

Copyright (C) 2026 vikash. See the COPYRIGHT file for details.

This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0). See the LICENSE file for details.