Container-orkestrering automatiserer udrulning, skalering, netværk og administration af containere på tværs af mange maskiner. Efterhånden som applikationer vokser til mange containere på tværs af flere værtsmaskiner, bliver manuel administration upraktisk — orkestratorer som Kubernetes håndterer det. At forstå behovet er vigtigt for at skalere containeriserede applikationer.
Problemet: administration af containere i stor skala
Running a few containers on one host is manageable. But in production at scale:
→ MANY containers across MANY hosts (dozens/hundreds of services and instances)
→ containers crash → need to restart automatically
→ load increases → need to scale up (more instances) / scale down
→ deployments → roll out new versions without downtime; roll back on failure
→ distribute containers across hosts; route traffic; service discovery
→ Doing all this MANUALLY is impractical and error-prone. → ORCHESTRATION automates it.
