Surviving DevOps: Key Practices to Eliminate Waste and Ensure Reliability

21CTO community guide: Whether you are just starting your development career, have a solid project portfolio, or are considering a shift to DevOps, this is my view on how to survive in this demanding world.

I have been reading the "Phoenix Plan" book and realized many patterns in my career; while I don’t aim to become a DevOps specialist, I see the importance of a DevOps team so developers don’t have to handle server deployment, configuration, or change management.

First Goal

Eliminate waste through automation, system monitoring, and a culture of information sharing across business, development, and operations.

This goal is crucial: IT operations and development must collaborate, not operate in silos, which is harder in practice than theory. The idea comes from Lean Six Sigma, which improves performance by systematically removing waste and variation.

You must communicate with both senior and junior management to identify and eliminate all interferences and waste.

Stability is paramount; maintain high standards or you’ll spend nights in the server room. You act as the glue binding business, development, and operations—if that glue thins, the business suffers.

Ensuring Coverage When Disasters Strike

Disasters are inevitable—whether caused by human error, malicious insiders, buggy software, or external factors. Complex systems mean failure.

Complex systems mean failure.

A complex system contains more than seven interacting nodes or components (e.g., UI, web app, backend service, database server, OS). Humans can only effectively manage a limited number of items simultaneously.

To meet these challenges, everything must be automated.

Run tests: all test cases, including unit tests, should be part of the pipeline to prevent interruptions on failures.

Collect metrics: gather logs and monitoring data in a non‑blocking way, storing them securely.

Run reports and alerts.

Self‑repair: software running on servers should adjust itself to recover from failures automatically.

Backup

This is the minimum requirement.

Record Everything

Automation: keep everything in automated scripts; maintain a real‑time wiki via APIs for configs, requirements, validation rules, etc.

Continuity: regenerate documentation on every change and keep it in sync with the current environment.

Traceability: documentation must include metadata—who updated, when, and why.

Detail: provide thorough information, not just vague statements like "updated 10 terminals with latest OS patches"; include patch versions, steps, and observed effects.

Control Change Requirements

Developers must ensure new code integrates seamlessly, while operators need tighter collaboration and automated deployment processes.

This leads to frequent releases.

Frequent releases are not a problem; they indicate a healthy pipeline. Keep track of business needs and development status to plan releases on time.

Conclusion

We have explored the expectations of the DevOps role from a developer’s perspective. Remember, this is just one viewpoint; as you define your own role, you will discover additional best practices.