Cloud-native development is a set of architectural principles and practices that allow applications to take full advantage of cloud infrastructure — not just its cost model but its scalability, resilience, and operational characteristics.
The twelve-factor app methodology — originally articulated by the Heroku team — remains the best concise statement of cloud-native principles. Twelve factors cover everything from configuration (stored in environment variables, not code), to disposability (processes start fast and shut down gracefully), to statelessness (application instances share no state; state lives in external data stores). Together they describe an application that can be deployed, scaled, and replaced at any time without data loss or availability impact.
Containerisation is the packaging standard of cloud-native applications. Docker containers provide consistent, reproducible deployments across development, test, and production environments. They eliminate "works on my machine" as a category of production incident.
Microservices architecture — decomposing applications into independently deployable services — enables teams to release features independently, scale services according to their specific load, and choose the best technology for each function.
Event-driven architecture decouples services through asynchronous messaging, increasing resilience and reducing the blast radius when individual services fail. A message queue absorbs traffic spikes and ensures no data is lost if a downstream service is temporarily unavailable.
Observability — logs, metrics, traces — is a first-class concern in cloud-native systems, not an operational afterthought. The complexity of distributed systems means that debugging without observability is like navigating a city without GPS.
For organisations building new systems today, cloud-native principles should be the default. Retrofitting them onto legacy applications is possible but expensive. The time to build cloud-native is at the start.