Trino Integration Overview

Trino serves as the primary query execution engine for the MATIH platform. The Query Engine integrates with Trino for complex analytics, federated queries across multiple data sources, and high-performance SQL processing. This section covers configuration, catalog management, data federation, and connector setup.

Architecture

The Query Engine communicates with Trino via its REST API and JDBC interface. Each tenant's queries are routed to Trino with tenant-specific session properties and catalog access controls.

Query Engine              Trino Coordinator            Trino Workers
     |                          |                          |
     |--- Submit query -------->|                          |
     |    (with session props)  |--- Parse & plan -------->|
     |                          |--- Distribute tasks ---->|
     |                          |                          |--- Execute
     |                          |                          |--- Return results
     |<-- Query results --------|<-- Aggregate results ----|

Key Features

Feature	Description
Federated queries	Query across Delta Lake, Iceberg, PostgreSQL, and more
Session properties	Per-tenant resource limits and configuration
Catalog isolation	Tenants access only their authorized catalogs
Query queueing	Resource groups manage concurrency per tenant
Cost-based optimization	Automatic query plan optimization

Trino Deployment

Trino is deployed as part of the data infrastructure via Helm:

Component	Replicas	Description
Coordinator	1	Query parsing, planning, and coordination
Workers	2+ (auto-scaled)	Distributed query execution

Section Pages

Page	Description
Configuration	Trino connection settings and resource groups
Catalogs	Catalog configuration for data sources
Data Federation	Cross-source query federation
Connectors	Available Trino connectors and setup

Trino Integration Configuration