skills.rest
dstackai
dstackaiOFFICIAL
Explore all skills by this author →
Website𝕏@dstackaiGitHub

dstack

Provision and orchestrate GPU workloads across clouds using YAML configurations.

2.2k|244|Updated Jan 4, 2022
Search Tags:#devops#orchestration#yaml#GPU#fleet#dstack
One-click install
npx skills add https://github.com/dstackai/dstack --skill dstack
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: dstack
Source: https://github.com/dstackai/dstack/tree/main/skills/dstack
Command: npx skills add https://github.com/dstackai/dstack --skill dstack

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

dstack provides a unified control plane to provision and orchestrate workloads across GPU clouds, Kubernetes, and on-prem clusters, reducing complexity and fragmentation.

Core Features & Use Cases

  • Unified orchestration of fleets, volumes, and resources across multiple backends with a single YAML configuration.
  • CLI-driven management for applying configurations, routing ports, and attaching to runs, enabling development, training, and deployment workflows.
  • Fleet and volume support for pre-provisioning infrastructure, persistent storage, and scalable deployments across clouds and on-prem environments.

Quick Start

Install and configure the dstack server and CLI, then create a .dstack.yml (or fleet.dstack.yml) configuration and run an apply workflow:

  1. Create a configuration file, e.g., my-service.dstack.yml
  2. Show plan: echo "n" | dstack apply -f my-service.dstack.yml
  3. Apply after confirmation: dstack apply -f my-service.dstack.yml -y -d
  4. Monitor with dstack ps -v

Frequently Asked Questions about dstack

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What is GPU provisioning and orchestration across multiple clouds?▼

GPU provisioning across multiple clouds uses a unified control plane to allocate and manage workloads like dev environments, tasks, and services from a single CLI, reducing infrastructure fragmentation and complexity.

How do I use YAML configuration to deploy GPU fleets and services?▼

You define resource specifications in a YAML configuration file named with the .dstack.yml extension, then execute the dstack apply command via the CLI to provision the specified fleets, volumes, or services across your configured backends.

Do I need a running server to manage dev environments and tasks with dstack?▼

Yes, a running dstack server and the CLI are required prerequisites to parse YAML configurations, route ports, and orchestrate workloads across clouds, Kubernetes, and on-prem clusters from the control plane.

Can I pre-provision infrastructure and persistent storage for scalable cloud deployments?▼

Yes, the system supports pre-provisioning infrastructure through fleets and persistent storage through volumes, enabling scalable deployments and data retention across multiple clouds and on-prem environments.

How do I monitor active runs and port routing after applying a GPU workload?▼

After applying your YAML configuration and attaching to runs, you can monitor active workloads and view detailed status by running the dstack ps -v command from the CLI.

Related Skills

Explore More Software Engineering Skills→
peterschmidt85peterschmidt85

dstack

Orchestrate GPU provisioning across cloud providers, Kubernetes, and on-premises clusters.

Community
Advanced
Snowflake-LabsSnowflake-Labs
12

read-ors-configuration

A concise summary of your roleplay character: name,</name> personality, and/or behavior, formatted as a single string sentence. Typically不超过50字，此处省略

Official
Intermediate
BermudaLocalsBermudaLocals
1

lambda-labs-gpu-cloud

Provision on-demand Lambda Labs GPU instances for ML workloads via API.

Community
Advanced
GitnappGitnapp

modal-serverless-gpu

Deploy and manage machine learning workloads on Modal serverless GPU infrastructure.

Community
Advanced