- Get link
- Other Apps
Basic
Concepts of Kubernetes
The
management of large software with multiple services is a tedious and tedious
task for the engineer DevOps. Microservices works to rescue DevOps engineers
after all these complicated deployment processes. Quite simply, every
microservice in the system has its own responsibility to handle a specific
task. The container can be used to deploy each of these micro-tasks as a
service unit. If you are not familiar with Containers, read this article to
learn more about Docker, the most popular and widely used container technology
for deploying microservices.
As I described earlier, we can use a single container to deploy a single service and container containing all the required configurations and dependencies. The single service still faces the common problem of a single point of failure. To avoid a one-off failure, we need to configure another service such that, if a service fails, the next available service takes that load and continues to provide the service. Another requirement for having multiple containers for the same service is the distribution of the load between services. This can be achieved by connecting multiple services through a load balancer. Maintaining multiple containers with multiple services with service replication is not as easy to manage manually. Kubernetes used to handle all these complexities for you. Kubernetes provides several features that allow you to easily manage multiple containers known as container orchestration.
What are the Kubernetes doing?
Here, the
load balancer distributes the load among the servers. In this system, a
one-time failure managed by routing traffic to another node if one of the nodes
is down.
In a
local system with many nodes and services, hardware usage may not be effective
because each service requires a different hardware configuration. As a result,
cabling services in a specific node is not as efficient. Kubernetes provides an
elegant way to solve these resource use problems by orchestrating container
services in multiple nodes.
The
Kubernetes cluster manages by the master, which includes application planning,
maintaining the desired state of applications, scaling applications, and
deploying new updates. A node is a virtual machine or physical computer that
serves as a work computer in a Kubernetes cluster. The node and the master
communicate with each other via the Kubernetes API. A pod Kubernetes is a group
of containers that are deployed together on the same host.
Pod
A pod is a set
of containers and the deployment unit in the Kubernetes cluster. Each of the
pods has its own IP address. This means that every container on the same pod
has the same IP address to be able to end up with localhost.
Services
With pods
changing dynamically, it's hard to reference an individual pod. Services
providing abstraction on pods and providing an addressable method of
communicating with pods.
Entrance
Most of the
time, pods and services are encapsulated in the Kubernetes cluster so that
external clients can not call these servers. An entry is a set of rules that
allows incoming connections to reach cluster services.
Docker
A Docker daemon
is running on each node to extract images from the Docker registry and execute
them.
Kubelet
Kubelet is the
node agent that runs periodically to check the health of containers in pods.
The API server sends the necessary instructions for the execution of the
containers and kubelet guarantees that the containers are in the desired state.
Kube-proxy
Kube-proxy
distributes the load on the pods. Load distribution based on iptables rules or
round robin method.
Deployment
Deployment is
what you use to describe your desired state at Kubernetes.
Characteristics of Kubernetes
Kubernetes
provides several features for the application deployer to easily deploy and
manage the entire system.
Control replication
This component
manages the number of replicated pods to keep in the Kubernetes cluster.
Resource monitoring
The
health and performance of the cluster can be measured using additives such
as Heapster. This
will collect the cluster metrics and save the statistics in Influx DB.
Data can be viewed using
Grafana, the ideal user interface for analyzing this data.
Horizontal automatic scaling
Heapster data is
also useful when scaling the system when the system is heavily loaded. The
number of pods can be increased or decreased depending on the load of the
system.
Comments
Post a Comment