← DevOps
DevOps practices

Practices.

A practical reference for the frameworks and working patterns that shape reliable delivery.

Practice group

Architecture & adoption

Frameworks for making cloud decisions deliberately and revisiting them as the system evolves.

Well-Architected Framework

A repeatable review of whether a workload is secure, reliable, efficient, cost-aware, sustainable, and operationally sound.

Use it when

Designing a new workload, preparing a production launch, or reviewing an existing platform.

Key ideas
  • Operational excellence
  • Security
  • Reliability
  • Performance efficiency
  • Cost optimization
  • Sustainability
Working loop
  1. Describe the workload and its business goals
  2. Review each pillar with evidence
  3. Record risks and trade-offs
  4. Prioritize remediation work
  5. Repeat after meaningful change

Cloud Adoption Framework

A structured path from cloud strategy to landing zones, migration, governance, security, and ongoing management.

Use it when

Starting or scaling cloud adoption across multiple teams, subscriptions, accounts, or workloads.

Key ideas
  • Strategy and outcomes
  • Adoption plan
  • Landing zone readiness
  • Migration and modernization
  • Governance
  • Management and security
Working loop
  1. Define motivations and measurable outcomes
  2. Inventory workloads and skills
  3. Prepare identity, networking, policy, and account structure
  4. Adopt in controlled waves
  5. Govern and operate continuously
Practice group

Delivery & source control

Practices that keep changes small, reviewable, reproducible, and safe to release.

Branching Strategy

An agreed model for how code moves from an idea to an integrated, releasable change.

Use it when

Defining repository rules or fixing slow integration, merge conflicts, and unclear release ownership.

Key ideas
  • Trunk-based development
  • GitHub Flow
  • GitFlow
  • Release branches
  • Short-lived feature branches
  • Protected branches
Working loop
  1. Start from release frequency and risk
  2. Choose the simplest fitting model
  3. Keep branches short-lived
  4. Automate checks and reviews
  5. Document hotfix and release paths

Continuous Integration & Delivery

Integrate frequently and use an automated path to build, test, package, and promote every change.

Use it when

Teams need faster feedback, consistent artifacts, and repeatable releases.

Key ideas
  • Fast feedback
  • Build once, promote many
  • Automated quality gates
  • Environment parity
  • Progressive delivery
Working loop
  1. Trigger on every change
  2. Run fast validation first
  3. Create an immutable artifact
  4. Promote with policy and approvals
  5. Observe the release and roll back safely

GitOps

Operate systems from declarative desired state stored in Git and reconciled automatically by controllers.

Use it when

Managing Kubernetes or other declarative platforms where auditability and drift correction matter.

Key ideas
  • Git as source of truth
  • Pull-based reconciliation
  • Declarative state
  • Drift detection
  • Auditable changes
Working loop
  1. Change desired state through a pull request
  2. Validate policy and configuration
  3. Merge the approved state
  4. Let a controller reconcile
  5. Alert on failed reconciliation or drift
Practice group

Operations, security & cost

Shared disciplines that turn reliability, security, and financial responsibility into daily engineering work.

Site Reliability Engineering

Apply software engineering to operations and balance reliability work with product delivery through explicit service objectives.

Use it when

A service needs measurable reliability targets and a disciplined response to operational load.

Key ideas
  • SLIs and SLOs
  • Error budgets
  • Toil reduction
  • Blameless postmortems
  • Capacity planning
Working loop
  1. Measure user-visible reliability
  2. Set realistic objectives
  3. Use error budgets to guide release risk
  4. Automate repetitive work
  5. Learn from incidents

DevSecOps

Make security a shared, automated part of design, development, delivery, and operations.

Use it when

Security feedback arrives too late or ownership is isolated in a separate approval stage.

Key ideas
  • Threat modeling
  • Least privilege
  • Software supply chain
  • Policy as code
  • Continuous scanning
Working loop
  1. Model threats early
  2. Protect identities and secrets
  3. Scan code, dependencies, images, and infrastructure
  4. Enforce risk-based policy
  5. Monitor and respond in production

Infrastructure as Code

Define infrastructure in versioned, reviewable configuration and apply it through automated workflows.

Use it when

Infrastructure must be reproducible across environments and changes need an audit trail.

Key ideas
  • Declarative configuration
  • Reusable modules
  • Remote state
  • Plan before apply
  • Drift management
Working loop
  1. Model the desired infrastructure
  2. Review a generated plan
  3. Apply through automation
  4. Verify outcomes
  5. Detect and resolve drift

FinOps

Create shared visibility and accountability so engineering, finance, and business teams optimize cloud value together.

Use it when

Cloud spend is growing, ownership is unclear, or cost decisions are disconnected from product value.

Key ideas
  • Allocation and tagging
  • Unit economics
  • Budgets and forecasts
  • Rightsizing
  • Commitment management
Working loop
  1. Make cost data visible
  2. Allocate spend to owners
  3. Define useful unit metrics
  4. Optimize architecture and usage
  5. Review forecasts and commitments continuously
Search