session-56.txt

everything is resource inside k8s
K8s -> PaaS

namespace -> an isolated space inside k8 cluster where can create our project related resources. we can completely control them. it is like our project inside k8s.

1. namespace level -> VPC level
2. cluster level -> global level

kind: <type-of-resource>
apiVersion: 

there are resources categorised based on apiVersion..

pod
===
pod is smallest deployable unit in kubernetes. a pod can have multiple containers in it. containers inside pod share same network and storage

docker run -d -p 80:80 --name frontend nginx:1.24

kubectl describe pod pod-name -n namespace

assign to node
node pulls the image
create container
start container

kubectl exec -it multi-container -c nginx -- bash

labels
======
lables adds metadata to the resource, labels are selectors inside kubernetes..

CrashLoopBackOff -> when container is not able to start
ImagePullBackOff/ErrorImagePull -> when your node is not able to pull the image.. image address may be wrong or authentication problems

annotations:
======
annotations are again like labels adds metadata, but annotations are used to select external resources..
labels have limitations in key value length and size. special charecters are not allowed inside labels

resources
=======
you need to always restrict the resources to the pod..if you get more requests we can autoscaling

1cpu = 1000m

env
====
env variables useful for pod

siva -> saiavaaa

service
=======
pod1 -> pod2

if we want pod to pod communication, ip address of pods are not useful since they are ephemeral. we can use k8s service to acheive
1. pod to pod communication through DNS, service name works as DNS here
2. load balancing

cart -> catalogue

pod1 -> service -> pod2, pod3, pod4

1. ClusterIP
2. NodePort
2. NodePort
3. LoadBalancer

service ip: 10.100.10.134
endpoints(pods): 192.168.5.23:80

svc-test -> service -> labels

url/repo-name/image-name:version