bobmatnyc/claude-mpm-skills

digitalocean-compute

DigitalOcean compute services covering Droplets, App Platform, Functions, Kubernetes (DOKS), GPU Droplets, and Bare Metal GPUs.

查看源码
仓库原始内容

按源仓库内容呈现,保留标题、案例、代码、表格、链接以及原文引用的演示图片。

DigitalOcean Compute Skill


progressivedisclosure: entrypoint: summary: "Compute on DigitalOcean: Droplets, App Platform, Functions, Kubernetes, GPU Droplets, Bare Metal GPUs." whentouse:

  • "When choosing a compute model for an application"
  • "When provisioning VMs, PaaS, serverless, or Kubernetes"
  • "When planning GPU-backed workloads"

quick_start:

  • "Pick Droplets, App Platform, Functions, or Kubernetes"
  • "Select region, size, and image"
  • "Configure networking and access"
  • "Deploy and monitor workloads"

token_estimate: entry: 90-110 full: 4000-5200


Overview

DigitalOcean compute includes VMs, managed PaaS, serverless functions, and managed Kubernetes, plus GPU offerings for AI workloads.

Droplets (VMs)

Use Droplets for full OS control and custom networking.

  • Choose region and Droplet plan.
  • Select an OS image or snapshot.
  • Add SSH keys for secure access.
  • Attach Volumes for persistent block storage.
  • Apply firewalls and VPC configuration.

App Platform (PaaS)

Use App Platform to deploy from Git or container images with managed build and runtime.

  • Connect a Git repository or container registry.
  • Define build and run commands or app spec.
  • Set environment variables and secrets.
  • Configure custom domains and HTTPS.

Functions (FaaS)

Use Functions for event-driven or lightweight API workloads.

  • Create a Functions namespace.
  • Deploy functions with doctl or the Control Panel.
  • Configure environment variables and triggers.
  • Monitor execution and logs.

Kubernetes (DOKS)

Use managed Kubernetes for multi-service container orchestration.

  • Create a cluster and node pools.
  • Save kubeconfig for cluster access.
  • Deploy workloads with Kubernetes manifests.
  • Pull images from Container Registry.

GPU Compute

Use GPU Droplets or Bare Metal GPUs for AI and ML workloads.

  • GPU Droplets: VM-based GPU workloads with flexible sizing.
  • Bare Metal GPUs: Dedicated GPU hardware for high performance.

Compute Selection Guide

  • Use Droplets for full control and custom configurations.
  • Use App Platform for managed app deployments from Git or containers.
  • Use Functions for event-driven workflows and micro APIs.
  • Use Kubernetes for multi-service container orchestration.
  • Use GPU Droplets or Bare Metal GPUs for AI or GPU-heavy workloads.

Complementary Skills

When using this skill, consider these related skills (if deployed):

  • digitalocean-containers-images: Registry and image workflows.
  • digitalocean-storage: Spaces, Volumes, and backups.
  • kubernetes: Cluster operations and troubleshooting.
  • docker: Container build and runtime patterns.

Note: Complementary skills are optional. This skill is fully functional without them.

Resources

DigitalOcean Docs:

  • Compute: https://docs.digitalocean.com/products/compute/
  • Droplets: https://docs.digitalocean.com/products/droplets/
  • App Platform: https://docs.digitalocean.com/products/app-platform/
  • Functions: https://docs.digitalocean.com/products/functions/
  • Kubernetes: https://docs.digitalocean.com/products/kubernetes/
  • GPU Droplets: https://docs.digitalocean.com/products/gpu-droplets/
  • Bare Metal GPUs: https://docs.digitalocean.com/products/bare-metal-gpus/
来自同一仓库

更多 Skills

全部 Skills