DevOps

What is it

DevOps combines the terms Development and Operations. It’s not just a tool or a methodology — it’s a culture of collaboration between those who build software and those who keep it running.

The goal: deliver software changes frequently, reliably, and automatically, breaking down the silos that historically separated developers and system administrators.

The Problem It Solves

In the traditional model:

Dev Team: "The code is ready, let Ops deploy it"
                     ↓
Ops Team: "We can't deploy until we review everything"
                     ↓
Result: Weeks of waiting, production bugs, blame games

With DevOps, both teams share responsibility and work together from the start.

The 5 pillars of CALMS

Principle	Description
Culture	Collaboration and shared responsibility
Automation	Automate everything repeatable
Lean	Eliminate waste from the process
Measurement	Measure everything to continuously improve
Sharing	Share knowledge across teams

The DORA Metrics

DORA (DevOps Research and Assessment) popularized four software delivery metrics for years; its official guide now uses five, grouped into throughput and instability:

Metric	Type	What it measures
Change lead time	Throughput	Time from commit to deployment in production
Deployment frequency	Throughput	How many deployments happen over a period
Failed deployment recovery time	Throughput	Time to recover from a failed deployment (replaces “Time to Restore Service”/MTTR)
Change fail rate	Instability	% of deployments that require immediate intervention
Deployment rework rate	Instability	% of unplanned deployments made to fix a production incident

Top performers do well across all five at once: speed and stability are not a trade-off.

Typical DevOps Ecosystem Tools

Code → [Git] → CI/CD Pipeline → Tests → Build → Deploy → Monitor
                (GitHub Actions,           (Docker,    (Datadog,
                 GitLab CI,             Kubernetes)   Grafana)
                 Jenkins)

Version Control: Git, GitHub, GitLab
CI/CD: GitHub Actions, GitLab CI, Jenkins, CircleCI
Containers: Docker, Podman
Orchestration: Kubernetes, Amazon ECS
Infrastructure as Code: Terraform, Ansible, Pulumi
Monitoring: Datadog, Grafana, Prometheus, New Relic

DevOps in Practice: A Modern Delivery Cycle

Developer commits → code goes to Git
CI pipeline triggers automatically
Tests run in parallel (unit, integration)
Build is packaged into a Docker image
Automatic deploy to staging environment
Smoke tests confirm it works
Approval (manual or automatic) for production
Deploy to production with automatic rollback if something fails
Continuous monitoring and alerts

In mature teams, this whole cycle can take minutes or hours instead of weeks.

Why It Matters for Companies Hiring Software

When evaluating a software vendor, DevOps practices signal team maturity:

Team with good DevOps practices:

Can deliver small, frequent changes (lower risk)
Catches problems before they reach production
Recovers quickly when something fails
Has full visibility into what’s happening in production

Team without DevOps:

Large, infrequent, risky deployments
Late bug discovery (when the customer already suffered it)
Slow, manual incident recovery
Little control over the real state of the system

CI/CD - The central DevOps practice: continuous integration and delivery
Docker - Containers that make repeatable deployments possible
Shift-Left Security - Bringing security to the start of the DevOps cycle
Microservices - Architecture that DevOps allows deploying independently

Additional Resources:

DORA Research - Research on DevOps performance metrics
The DevOps Handbook - Reference book

Pronunciation