This is an open source container orchestration framework (coordination and management of multiple containers)
Helps manage containers → Kubernettes helps manage applications which are made up of Containers, and helps you manage them in different environments (Cloud, Physical machines, VMs, or even Hybrid envs)
What problems does K8s solve?
-
From Monolith to Microservers
-
Increased usage of containers
-
Rise of containers → we need a container management tool to manage them across multiple environments
Features of K8s
- High availability (Application has no downtime!), and scale up and down as needed
- Disaster recovery → Backup and restore
Kubernetes Architecture
- 1 master node (Virtual or Physical Machine) which is connected by worker nodes
- Each worker node has a Kubelet node on it (Kubernettes process which allows the clusters to communicate w/ each other, and execute tasks, etc)
- Each worker node has containers deployed on it (of different applications)
- Worker node is where the applications are running (containerized microservices)
- Worker nodes have higher workload, and much bigger and more resources (as they run the apps)
- Master node → Runs several k8s processes: API, Controller Manager, and Scheduler
- Master Node → Control Plane Nodes (Handful of master process, but much more important as it has control to everything!)
- Usually we run multiple copies of master nodes incase we lose 1 or 2, we can still manager our containers with the replacement
- API server, entry point to the kubernettes cluster → This is the process which different clients can talk to i.e. UI, APIs, CLI
- Controller Manager → Keeps track of whats happening in the cluster (if something needs to be repaired, if a container died and needs to be restarted.. etc)
- Scheduler → Ensures Pods placements (Responsible for scheduling containers on different nodes based on capacity, and load)
- ETCD → Kubernettes backing store (Contains the Current State of the K8s at any time → has configuration data of and state of every worker node
- Also used for backup and restore by snapshotting all states from this etcd
- Virtual Network → Allows the communication between master and worker nodes → Creates one unified machine
Kubernetes Components — There is alot!
- We will look at each component w/ the help of building a simple webapp, with a database
- Worker Node / Node → VM or physical server
- Application Pod → Smallest unit in Kubernetes, which offers abstraction over container. This creates a running env on top of the container → This allows us to switch away to any container (don’t have to be tied to just Docker, … etc)
- Usually 1 app / pod
- pod can be a Database on the same worker node
- Each pod gets it’s own IP address, which allows us to communicate w/ each other using that IP address
- New IP address on creation
- Pod → Layer on top of the container → Abstraction over container → Allows us to replace containers,