Cloud#
For the operator, Cloud is the ability to build and deploy infrastructure defined in code. The discipline at the center is Infrastructure as Code: a Git repository describes what the cloud should look like, a pipeline applies the description, and the cloud reconciles itself to match.
In industry terms, this is the DevOps and platform engineering discipline: the same operator skills, the same tooling, often the same team. SRE shares most of the same stack with a stronger emphasis on reliability and on-call.
Concretely, the section covers configuration management (Puppet, Chef, Ansible), Infrastructure as Code (Terraform, Pulumi, CloudFormation, CDK), cloud providers (AWS, GCP, Azure), the components those providers expose (compute, networking, storage, security), VM and host provisioning automation, application deployment, the CI/CD and GitOps pipelines that ship them, and the observability stack (metrics, monitoring, alerting, logs, and traces) that keeps the system visible in production.
The seven sub-trees below cover the layers around that discipline,
what the operator deploys (runtimes), how it ships (pipelines),
where it runs (clouds), how it’s described (iac), what it’s made of
(components), how it’s shaped (architecture), and how it’s operated
(practices). Bare-metal, VM,
and traditional-ops material is kept as prerequisite context in
runtimes/; most modern stacks still have a layer of it
underneath the cloud.
Culture, on-call rotation, and release flow that tie code, pipeline, and production together.
Bare metal, VMs, containers, and orchestration. The substrate everything else runs on.
CI/CD and GitOps. How code turns into a deployed artifact.
AWS, GCP, Azure, DigitalOcean, Cloudflare, and the smaller providers; their primitives and their gotchas.
Terraform, Pulumi, CloudFormation, CDK. The Git-tracked description of what the cloud should look like.
Compute, networking, storage, and security primitives the cloud exposes.
Landing zones, network topology, resilience, tenancy, and cost. How the components are composed into one system.
Puppet, Chef, Ansible, and friends. Driving in-host state.
Metrics, logs, and traces emitted from the application and its dependencies.
Dashboards, alerting, and the on-call workflow that turns signals into action.
SRE, platform engineering, and the operating disciplines that hold the stack together.