Engineering ReferenceEngineering Reference

DevOps

On this page

DevOps unifies software development and IT operations. The goal: ship changes faster, more reliably, with shorter feedback loops — by automating builds, tests, deployments, and operations.

Overview

The reference metrics come from DORA: deployment frequency, lead time for change, change failure rate, mean time to restore. Elite performers deploy on demand, in under an hour, with <15% failures, and recover in <1 hour.

Pillars

  • Culture — shared ownership; no “throw over the wall”.
  • Automation — everything that can be a script.
  • Lean flow — small batches, trunk-based development.
  • Measurement — DORA, SLOs, error budgets.
  • Sharing — internal docs, runbooks, blameless postmortems.

CI/CD

  • CI — every commit builds, tests, lints, packages.
  • CD — every green build is potentially deployable.
  • Strategies: blue/green, canary, feature flags, dark launches.
  • Artifact registries: Docker Hub, ECR, Artifactory, GitHub Packages.
  • Pipelines: GitHub Actions, GitLab CI, Jenkins, CircleCI, Argo CD, Tekton.

Infrastructure as Code

  • Terraform / OpenTofu — cloud-agnostic resource provisioning.
  • Pulumi — IaC in real programming languages.
  • CloudFormation, Bicep, ARM — vendor-specific.
  • Ansible, Chef, Puppet, Salt — config management.
  • Helm, Kustomize — Kubernetes packaging.
  • State management & remote backends are non-optional.

Observability

  • Three pillars: metrics, logs, traces.
  • OpenTelemetry for instrumentation.
  • SLO/SLI/SLA: define what “working” means; alert on burn rate.
  • Tools: Prometheus + Grafana, Datadog, New Relic, Dynatrace, Honeycomb, Splunk, ELK/EFK.

Tools

  • Version control: Git, GitHub, GitLab, Bitbucket.
  • Containers: Docker, Podman, BuildKit.
  • Orchestration: Kubernetes, ECS, Nomad.
  • Secrets: Vault, AWS Secrets Manager, SOPS.
  • Incident: PagerDuty, Opsgenie, Statuspage.
reference page