DigitalOcean

DigitalOcean#

DigitalOcean is the developer-first public cloud. Smaller catalog than the hyperscalers, simpler pricing, fewer knobs. The operator picks it for small teams, side-projects with a real production posture, and workloads where the hyperscaler’s complexity is the dominant cost.

Account model#

  • Team, the billing and access unit.

  • Project, in-team grouping for resources.

  • Member, a user with a role (Owner, Member, Billing).

  • API tokens for programmatic access; one personal access token per team or per machine identity.

Catalog at a glance#

Layer

Headline services

Compute

Droplets (VMs), DigitalOcean Kubernetes (DOKS), App Platform (PaaS), Functions (FaaS).

Storage

Spaces (object, S3-compatible), Volumes (block), Snapshots, Backups.

Database

Managed Databases for PostgreSQL / MySQL / Redis / Kafka / MongoDB, Caching.

Networking

VPC, Load Balancers, Cloud Firewalls, Floating IPs, Reserved IPs, DigitalOcean DNS.

Observability

Monitoring (basic metrics + alerts), built-in graphs.

Misc

Container Registry, Premium GPU droplets (Paperspace-derived), 1-Click Apps.

CLI and SDKs#

The doctl CLI plus the godo SDK for Go and community libraries elsewhere.

$ doctl auth init                          # API token at session start
$ doctl compute droplet list
$ doctl kubernetes cluster list
$ doctl spaces ls space://my-bucket
$ doctl apps create --spec spec.yaml

Regions#

NYC1/2/3, SFO2/3, AMS3, FRA1, LON1, SGP1, TOR1, BLR1, SYD1. No formal multi-AZ inside a region; the operator architects multi-region for HA when needed.

Pricing model#

The selling point. Flat per-resource pricing, predictable to the dollar:

  • Droplets from $4/month (1 vCPU, 512 MB) up through GPU droplets at the high end.

  • Spaces flat $5/month for 250 GB plus $0.01/GB egress over 1 TB.

  • Managed databases from $15/month.

No reserved instances, no spot, no committed-use discounts. What you see on the price page is what shows up on the bill.

When to pick DigitalOcean#

  • Side-projects, prototypes, small SaaS.

  • Teams that want the ergonomics of “just give me a VM” with a modern UI.

  • Workloads where the predictable bill matters more than catalog depth.

When to pick something else#

  • Anything requiring a service the hyperscalers offer and DO does not (deep ML, managed streaming, mainframe-adjacent, etc.).

  • Workloads needing strict compliance regimes (SOC 2 Type 2 only on Premium plan; no FedRAMP, no HIPAA BAA by default).

  • Multi-region active-active with a low-latency global front (use Cloudflare or a hyperscaler).

Datacenter locations#

Code

Location

nyc1

New York City, USA

nyc2

New York City, USA

nyc3

New York City, USA

sfo2

San Francisco, USA

sfo3

San Francisco, USA

atl1

Atlanta, USA

ric1

Richmond, USA

tor1

Toronto, Canada

ams3

Amsterdam, Netherlands

lon1

London, United Kingdom

fra1

Frankfurt, Germany

blr1

Bangalore, India

sgp1

Singapore

syd1

Sydney, Australia

References#

Contents