Glossary · Automation software engineering and architecture
Container orchestration
Also known as: Orchestration
German: Container-Orchestrierung
In software engineering, container orchestration is the automated deployment, scheduling, scaling, networking, health monitoring and recovery of containers across one or more hosts, performed by platforms such as Kubernetes or lightweight variants for edge devices.
- Software engineering
In one sentence
Container orchestration automates deploying, scaling, networking and recovering containers across hosts, for example with Kubernetes.
Example
An orchestration platform restarts a crashed data-collection container on the line's edge cluster automatically and moves it to another node if a host fails.
How it applies
- Engineering: Orchestration platforms work declaratively: operators describe the desired state, such as which images, how many replicas and which resources, and the platform works to maintain it. Configuration is usually stored as Infrastructure as code (IaC).
- Operation: Automatic restarts and rescheduling improve availability of IT and edge services. For shop-floor use, consider network outages, limited hardware, update windows and whether automatic actions could interfere with production.
- Documentation: Document the cluster architecture, node roles, deployed workloads, update and rollback procedures and how to diagnose failures. Operators need clear guidance on which actions the platform takes automatically and which require people.
Orchestration vs. runtime
A Container runtime starts and stops containers on one host. Orchestration coordinates many containers across hosts, using Health check and Liveness probe results to decide when to restart or replace them.
Orchestration does not by itself make applications deterministic or suitable for safety-related control.