Skip to content

Core

Main K8s Cluster Architecture

  • Main Node:
  • etcd: A distributed key-value store that stores the cluster's configuration data and provides a reliable source of truth.
  • kube-scheduler: Assigns Pods to nodes based on resource availability and other constraints.
  • kube-controller-manager: runs controller processes that loop to regulate the state of the system to achieve the desired state.

  • kube-api server: Exposes the Kubernetes API and acts as the front end for the Kubernetes control plane.

  • Worker Nodes:

  • kubelet: An agent that runs on each worker node and ensures that containers are running in a Pod.
  • kube-proxy: Maintains network rules on nodes, enabling communication between Pods and external traffic.
graph TD
  subgraph Main_Node
    kube_api_server -->|Exposes API| controllers
    kube_api_server -->|Exposes API| kube_scheduler
    kube_api_server -->|Exposes API| etcd
  end

  subgraph Worker_Node_1
    kube_api_server -->|Exposes API| kubelet1
    kube_api_server -->|Exposes API| kube_proxy1
    kubelet1 -->|Manages| Pod1.1

    kube_proxy1 -->|Manages| Pod1.1 
  end

  subgraph Worker_Node_2
    kube_api_server -->|Exposes API| kubelet2
    kube_api_server -->|Exposes API| kube_proxy2
    kubelet2 -->|Manages| Pod2.1
    kubelet2 -->|Manages| Pod2.2

    kube_proxy2 -->|Manages| Pod2.1
    kube_proxy2 -->|Manages| Pod2.2
  end

Main K8s Resources

  • Pods: The smallest deployable units in Kubernetes, consisting of one or more containers and shared resources, enabling the deployment of applications.
  • ReplicaSets: Kubernetes controllers that ensure a specified number of pod replicas are running, providing scalability, fault tolerance, and the ability to manage multiple identical instances.
  • Deployment: A higher-level abstraction managing ReplicaSets, enabling declarative updates, rollbacks, and the definition of desired application states in Kubernetes.
graph TD
  subgraph Cluster
    subgraph Node1
      pod1.1
      pod1.2
      pod1.3
    end

    subgraph Node2
      pod2.1
      pod2.2
      pod3.3
    end

    subgraph ReplicaSet
      replicaset1 -->|Creates| Node1
      replicaset1 -->|Creates| Node2
    end

    subgraph Deployment
      deployment -->|Manages| ReplicaSet
    end
  end
  • Services: Abstractions that define logical sets of pods and policies for accessing them, facilitating communication within the cluster or from external sources, with types such as NodePort, ClusterIP, and LoadBalancer.
  • NodePort: Exposes a service on a static port on each cluster node, enabling external access to the service.
  • ClusterIP: Exposes a service on a cluster-internal IP address, allowing communication only within the cluster.
  • LoadBalancer: Exposes a service externally using a cloud provider's load balancer, distributing incoming traffic across multiple pods for scalability and availability.
graph TD
  NodePort1 --> Node1
  NodePort2 --> Node2
  subgraph Cluster

    subgraph ClusterIP_Services
      db_svc <----> Node1
      api_svc <----> Node2
      db_svc <----> api_svc
    end
    subgraph Node1
      dbpod1.1
      dbpod1.2
      dbpod1.3
    end
    subgraph Node2
      apipod2.1
      apipod2.2
      apipod3.3
    end
  end
  subgraph LoadBalancerService
    LoadBalancer --> Node1
    LoadBalancer --> Node2
  end
  • Namespaces: Virtual clusters within a Kubernetes cluster, providing a way to partition and isolate resources, enabling multiple teams or projects to share the same cluster without interference.
graph TD
  subgraph Cluster
    subgraph Namespace1
      pod1
      pod2
      service1 -->|In Namespace1| pod1
    end

    subgraph Namespace2
      pod3
      pod4
      service2 -->|In Namespace2| pod3
    end

    subgraph Namespace3
      pod5
      pod6
      service3 -->|In Namespace3| pod5
    end
  end