DevOps Principles#

DevOps is not a tool, a job title, or a team — it is a set of practices and cultural philosophies that unify software development (Dev) and IT operations (Ops). The goal is to shorten the systems development life cycle while delivering features, fixes, and updates frequently and reliably.

The Three Ways of DevOps#

The foundational principles come from The Phoenix Project and The DevOps Handbook:

DevOps Culture — CALMS#

The CALMS framework describes the cultural pillars of DevOps:

DORA Metrics#

The DevOps Research and Assessment (DORA) team identified four key metrics that predict software delivery performance:

Why DORA Matters#

• These metrics are outcome-based, not activity-based
• They correlate with organisational performance (profitability, market share)
• They apply regardless of tech stack, team size, or industry
• They expose bottlenecks and guide improvement efforts

The Fifth Metric: Reliability#

In 2021, DORA added reliability (meeting or exceeding reliability targets) as a fifth metric, recognising that speed without stability is counterproductive.

Shift-Left#

“Shift-left” means moving activities earlier in the development lifecycle:

graph LR
    A[Plan] --> B[Code] --> C[Build] --> D[Test] --> E[Release] --> F[Deploy] --> G[Operate]
    
    style A fill:#e3f2fd
    style B fill:#e3f2fd
    style C fill:#e3f2fd
    style D fill:#e8f5e1
    style E fill:#e8f5e1
    style F fill:#fff3e0
    style G fill:#fff3e0

Benefits of Shift-Left#

• Bugs found earlier are cheaper to fix (10x–100x cost multiplier)
• Security vulnerabilities caught before they reach production
• Faster feedback loops for developers
• Reduced rework and context switching

Feedback Loops#

Fast feedback is the engine of continuous improvement. Every stage of delivery should produce signals:

graph TD
    A[Developer commits code] --> B[CI runs tests in seconds]
    B --> C{Pass?}
    C -->|Yes| D[Deploy to staging]
    C -->|No| E[Immediate notification]
    D --> F[Automated smoke tests]
    F --> G[Monitoring & alerting]
    G --> H[User feedback & metrics]
    H --> A
    E --> A

Collaboration Patterns#

Breaking Down Silos#

Blameless Postmortems#

When incidents occur, the focus is on systems not people:

1. Timeline — What happened, when?
2. Contributing factors — What conditions allowed this?
3. Impact — Who was affected and how?
4. Mitigation — What stopped the bleeding?
5. Action items — What systemic changes prevent recurrence?
6. Learnings — What did we discover?

DevOps vs SRE vs Platform Engineering#

These disciplines overlap but have distinct focuses:

How They Complement Each Other#

graph TD
    A[Platform Engineering] -->|Provides self-service tools| B[Development Teams]
    B -->|Apply DevOps practices| C[Continuous Delivery]
    C -->|Monitored by| D[SRE]
    D -->|Feeds reliability requirements back to| A

• DevOps provides the cultural foundation
• SRE implements reliability practices with engineering rigour
• Platform Engineering builds the tooling that makes DevOps practices accessible at scale

The Infinite Loop#

DevOps is often visualised as an infinity loop (∞) representing the continuous nature of the process:

Anti-Patterns#

Key Takeaways#

• DevOps is a cultural movement that prioritises flow, feedback, and learning
• The DORA metrics (deployment frequency, lead time, change failure rate, MTTR) are the gold standard for measuring delivery performance
• Shift-left moves quality, security, and testing earlier — where fixes are cheapest
• Fast feedback loops at every stage are what make continuous improvement possible
• DevOps, SRE, and Platform Engineering are complementary disciplines, not competitors
• Start with culture and measurement; tools follow naturally