Replication Loop

• Replica Set and replication controllers are based on the concept called replication loop.
• Replica set and that application controller start continuously monitoring the pods under its control by first finding the pod that matches its pod labels.
• Then it counts and checks the number of running pods matching the desired number of pods defining the YAML configuration.
• If any of these controllers notice any differences in the number of pods s these controllers adjust the parts accordingly.
• For example, if the controller has to maintain 3 pods and 2 pods is only running then it creates a new port in a healthy worker node to match the desired number of parts defined in the YAML file.
• Another scenario would be if there are too many Pods are running inside the cluster, then the controller removes the pod so that it matches the desired number of pods.
• So the ultimate aim here is these controllers will make sure that desired number of pods always running mentioned in the template matches.

Also, check

Kubernetes Tutorials