Skip to main content

Cloud for Trino Product Overview

For your information

The product is in public preview. During the public preview period:

  • the product is provided free of charge;
  • no SLA applies;
  • service availability is not guaranteed;
  • financial compensation for downtime or failures is not provided.

We will notify you of any changes to the terms in advance.

Cloud for Trino is a service for deploying and managing clusters with the Trino distributed analytical SQL engine on the Selectel Cloud Platform. Trino allows you to run distributed SQL queries against large volumes of data stored in various formats and sources. Learn more about the Trino tool in the official Trino documentation.

Locations where you can create Trino clusters are listed in the Cloud for Trino availability by location guide.

To work with Trino clusters, user types and roles, projects, and project limits and quotas are supported.

Versions

Trino clusters support versions 482 and 483.

Use Cases

  • analytical processing of large volumes of data;
  • scaling analytical computations through distributed in-cluster processing;
  • executing interactive ad-hoc SQL queries;
  • executing federated SQL queries across various data sources: databases, message brokers, S3, and others;
  • processing data in various formats, including lakehouse formats: Iceberg, Delta Lake, and Hudi;
  • building ETL/ELT pipelines: loading and transforming data between sources;
  • preparing data for BI tools: Redash, Superset, and others;
  • building data platforms together with cloud databases, S3, Cloud for Redash, Cloud for Hive Metastore.

Cluster Components

A Trino cluster consists of a coordinator and workers.

Coordinator

The coordinator is the management node of the cluster that:

  • receives requests from users;
  • plans SQL query execution;
  • distributes tasks among workers;
  • collects results from workers.

A Trino cluster always has only one coordinator.

Workers

Workers are worker nodes of the cluster that:

  • execute tasks from the coordinator;
  • read data from sources using connectors;
  • perform data operations;
  • exchange intermediate results with each other;
  • return final results to the coordinator.

A Trino cluster can have up to 10 workers. You can:

  • set a fixed number of workers;
  • or configure autoscaling — the number of workers will change automatically only within the specified range depending on the cluster load.

Connectors

A connector is an interface for accessing a data source of a specific type. It converts data from the source into Trino schemas, tables, columns, rows, and data types. This allows you to work identically with all data sources, regardless of their specific requirements.

The connector is selected when creating a catalog.

The following connectors are available in Trino clusters:

  • PostgreSQL;
  • MySQL;
  • Iceberg;
  • ClickHouse®.

Catalogs

A catalog is a configuration of parameters for connecting to a specific data source. Each catalog defines the data source type using a connector and contains connection parameters. You can select only one connector for each catalog.

A Trino cluster supports querying data from different catalogs within a single SQL query.

How it works

To get started, create a cluster in the Control panel.

When creating a cluster, you choose configurations for the coordinator and workers. Fixed configurations with a balanced vCPU:RAM resource distribution are available. All workers in the cluster will have the same configuration. If these configurations do not fit your needs, you can request a custom configuration. To do this, create a ticket and specify the resource ratio.

A Trino cluster can only be connected to a private subnet. You can enable external access to connect to the Trino cluster from the internet. Learn more in the Connect to a Trino cluster guide.

External access affects only connections to the cluster. By default, the Trino cluster itself can connect to external data sources from catalogs over the internet — no additional configuration is required.

Catalogs can be added when creating a cluster or in an existing cluster.

How to work with a Trino cluster

After creating a cluster, you can:

Areas of responsibility

Selectel provides

  • hardware selection for high performance of Trino clusters;
  • installation and standard configuration of Trino;
  • updates and maintenance of the operating system and utility software;
  • configuration and maintenance of the utility network in the cluster;
  • resource availability for cluster scaling if you initiated the scaling;
  • technical support.

Selectel is not responsible for

  • the correctness of catalog configurations and data source connection parameters;
  • availability and operability of external data sources;
  • having the necessary access permissions to data sources;
  • the efficiency of written SQL queries;
  • data schema and structure in the sources;
  • modifying the Trino cluster configuration according to the load profile;
  • initiating cluster scaling.

Pricing

During the public preview period, Cloud for Trino is provided free of charge. In the future, the service will become paid — we will notify you of any changes in terms in advance.

Limits and restrictions

Selectel sets account limits, project limits, and resource quotas. Learn more in the Quotas and limits guide.

In Trino clusters, only fixed configurations can be selected for the coordinator and workers; custom selection of vCPU and RAM is not available. If these configurations do not fit your needs, you can request a custom configuration. To do this, create a ticket and specify the resource ratio.

When creating a Trino cluster, you can add up to 10 catalogs. If you need more than 10 catalogs, add additional catalogs after creating the cluster; learn more in the Manage catalogs in a Trino cluster guide.

ClickHouse® is a registered trademark of ClickHouse, Inc. https://clickhouse.com.