balancing • HA • Health-checks • Scheduling based on resource requirements • Volumes and configurations mounting • Central logging • Basic resource monitoring • Better resource utilisation • Pluggable parts
with server configuration • Expose Service publicly ◦ through API server ◦ externalIPs ◦ nodePort (single node/all nodes) ◦ loadBalancer • Give them public url • Make it scale with a deployment • Rolling update of all replicas while live
container's stdout preferably in json • Test and know your resource requirements • Be aware of node sizes in the cluster • Stateless please • If not stateless then clustered please • Must handle unexpected immediate restarts