---
title: "Managed Machines"
canonical: https://trickest.com/docs/key-concepts/managed-machines
description: "Use Trickest-operated workflow compute and understand capacity, queues, and run limits."
---

# Managed Machines

## Overview

Managed machines are workflow workers provisioned and operated by Trickest. Choose an available managed fleet when running a workflow; you do not install the execution agent on your own hosts.

Managed workflow compute is available through Enterprise configurations. [Community](/docs/key-concepts/community-plan) workflows use [self-hosted machines](/docs/key-concepts/self-hosted-machines).

## How Managed Fleets Work

Trickest provisions machines as workflows need capacity. The **Managed compute** section under **Settings > Infrastructure > Fleet** shows active machines, machines terminating, the machine limit, utilization, and active runs.

Provisioning and shutdown take time. A pending job does not necessarily indicate a failure: it may be waiting for an upstream result or for capacity. Use the run's task details to distinguish these states.

## Scaling

The run's machine count is an upper bound on concurrency, not the number of workers guaranteed to be active throughout the run. A graph with one ready task cannot use ten machines for that task simply because the limit is ten.

Parallel branches and distributed nodes can create more ready tasks. The platform schedules them within the run and fleet limits. Reducing the run limit can keep a large run from consuming as much shared capacity.

## Fleet Limits and Sharing

Multiple runs can use the same fleet. They compete for available capacity; starting another run does not reserve a separate pool. Jobs queue when more work is ready than the fleet can execute.

Before raising a limit, check whether the workflow has parallel work to do and whether other active runs are using the fleet. Inspect output quality as well as speed when changing distribution.

## Next Steps

[Distributing and Scaling Jobs](/docs/using-the-app/workflow-and-executions/distributing-and-scaling-jobs) explains file splitting, folder distribution, and task inspection.

---
_Markdown source of https://trickest.com/docs/key-concepts/managed-machines._