DevOps Tutorial · DevOps Basics · Lesson 3 of 25

DevOps History & Milestones

How DevOps grew out of agile and the "wall" between developers and operations, the tools and books that shaped it, from Git and Jenkins to Docker, Kubernetes and GitHub Actions, and what modern DevOps looks like today.

Create a free account to save this page, take notes and track your progress. Sign in

How DevOps started

In the 2000s most companies split development (write features, change things) from operations (keep things stable, avoid change). Agile made development faster, but releases still went through slow, manual hand-overs.

In 2008 Andrew Shafer and Patrick Debois discussed "agile infrastructure" at the Agile conference. In June 2009, John Allspaw and Paul Hammond of Flickr gave the Velocity talk "10+ Deploys per Day: Dev and Ops Cooperation at Flickr". Inspired by it, Debois organised the first DevOpsDays in Ghent in October 2009, and the hashtag #devops named the movement.

DevOps was never one tool: it is a set of practices (automation, small batches, shared ownership, measurement) that the tools below made practical.

Timeline of milestones

  1. GitHub ActionsHistoric
    • CI/CD built into GitHub; OpenTelemetry formed the same year from OpenTracing and OpenCensus
  2. AccelerateHistoric
    • Forsgren, Humble and Kim publish the research behind the four DORA metrics
  3. Google SRE bookHistoric
    • Site Reliability Engineering popularises SLOs, error budgets and blameless postmortems
  4. Kubernetes 1.0Historic
    • Released with the founding of the Cloud Native Computing Foundation (CNCF)
  5. Terraform 0.1Historic
    • Infrastructure as code across clouds (see the Terraform tutorial)
  6. Kubernetes announcedHistoric
    • Google open-sources a container orchestrator based on its internal Borg experience
  7. DockerHistoric
    • Containers become easy to build, ship and run; The Phoenix Project novel is published the same year
  8. AnsibleHistoric
    • Agentless configuration management over SSH (after Puppet, 2005, and Chef, 2009)
  9. JenkinsHistoric
    • Forked from Hudson; becomes the most widely used CI server of the decade
  10. Continuous DeliveryHistoric
    • Jez Humble and David Farley's book defines the deployment pipeline
  11. DevOpsDaysHistoric
    • The first DevOpsDays in Ghent gives the movement its name
  12. 10+ deploys per dayHistoric
    • Flickr's Velocity talk on development and operations working together
  13. Amazon EC2Historic
    • Servers on demand by API: infrastructure becomes programmable
  14. GitHistoric
    • Linus Torvalds creates Git for Linux kernel development
  15. Agile ManifestoHistoric
    • Short iterations and working software, the starting point DevOps extended to operations

Modernisation turning points

  • Programmable infrastructure (2006 onwards): cloud APIs made servers disposable and scripts the way to create them.
  • Configuration management (2005-2012): Puppet, Chef and Ansible described servers as code.
  • Continuous delivery (2010): every commit goes through an automated pipeline that can release it.
  • Containers (2013) and orchestration (2014-2015): immutable images and declarative clusters replaced hand-tended servers.
  • Measurement (2014-2018): the DORA research showed that speed and stability improve together.
  • Supply-chain security (2020s): after high-profile build-system compromises and dependency vulnerabilities, signed artifacts, SBOMs and frameworks such as SLSA became part of the pipeline.
  • Platform engineering (2020s): internal platforms give developers self-service "golden paths" so every team doesn't rebuild the same pipeline.

How operations practice changed

Manual era Configuration management Containers and platforms
Servers Hand-built "pets" Scripted, long-lived Immutable images, replaced not patched
Releases Quarterly, manual Scripted, weekly Automated, many per day
Environments Drift apart Converged by agents Recreated from code
Rollback Restore backups Re-run old config Redeploy the previous image

What to adopt today

  • Trunk-based development with a CI pipeline under 10 minutes, and build once, deploy many.
  • Containers with an orchestrator or a managed container platform, and infrastructure as code.
  • Progressive delivery (canary or blue-green) with automated rollback.
  • SLOs with burn-rate alerting, structured logs, metrics and traces through OpenTelemetry.
  • Workload identity instead of stored keys, signed artifacts and dependency scanning.

Official resources

Last reviewed 21 October 2026. Release and support dates come from the official release pages linked above.

Was this page helpful?