Skip to content

architecture.md

Latest commit

 

History

History
92 lines (65 loc) · 2.86 KB

architecture.md

File metadata and controls

92 lines (65 loc) · 2.86 KB

Search Architecture

Overview

The Search service is a Kubernetes-native API built on the aggregated API server framework that provides advanced resource discovery capabilities through field filtering and full-text search. It enables platform users to efficiently query and locate resources across the cluster using powerful indexing and real-time event processing.

Architecture Diagram

Note

Below is a C4 container diagram of the service and it's dependencies. This is meant to model individual components in the system and their responsibilities. It does not aim to provide visibility into external system components that may be a dependency of this system.

Search service component software architecture diagram

Components

Search API Server

Purpose: Expose search capabilities as native Kubernetes APIs

Responsibilities:

  • Register custom API endpoints under search.miloapis.com/v1alpha1
  • Handle authentication and authorization via Kubernetes RBAC
  • Provide RESTful API for search queries
  • Manage custom resource definitions for the search service

Query Types:

  • Field Filtering: Exact match, prefix, range queries on structured fields
  • Full-Text Search: Fuzzy matching, phrase queries, relevance scoring

Resource Indexer

Purpose: Real-time indexing of platform resources from audit logs

Responsibilities:

  • Subscribe to NATS JetStream audit log topic
  • Filter events based on active index policies
  • Evaluate CEL expressions for conditional indexing
  • Extract and transform resource data for indexing
  • Manage documents in index backend with proper error handling and retries
  • Manage index lifecycle (creation, updates, deletion)
  • Bootstrap indexes from existing state

See the Resource Indexer Architecture for detailed design documentation.

Controller Manager

Purpose: Manages and validates resources for the search service

Responsibilities:

  • Validates and activates index policies

Index Backend Storage

Purpose: High-performance full-text search and indexing

Responsibilities:

  • Structured metadata (namespace, name, labels, annotations) filtering
  • Full-text searchable content

Note

We're targeting Meilisearch as our first integration backend for indexed storage.

etcd

Purpose: Distributed key-value store that provides reliable data storage needs

Responsibilities:

  • Store control plane resources for the Search API server (e.g. index policies)