The first 2 lines of every configuration file tell us what we want to create, like for deployments it will be like
apiVersion: apps/v1
kind: Deployment
For service it will be
apiVersion: v1
kind: Service
Every configuration file in kubernetes ahs 3 parts.
- metadata: Contains metadata for the configuration like, name and labels
- spec: contains various attributes depending on the ind of component you are creating, like for deployment it will contain
replicas,templateetc - status: this part is maintained by kubernetes. Whenever we make a change to the configuration file kubernetes checks if desirec state != actual state, and then make the changes to actual state. The data about actual state is given by etcd (brain of master node).
example of a conf file
apiVersion: apps/v1
kind: Deployment
metadata:
name: nginx-depl
spec:
replicas: 2
selector:
matchLabels:
app: nginx-depl
template:
metadata:
labels:
app: nginx-depl
spec:
containers:
- name: nginx-depl
image: nginx
resources:
limits:
memory: "128Mi"
cpu: "500m"
ports:
- containerPort: 80
Now the question is where inside the configuration file do we specify the configuration/blueprint of a pod inside a deployment configuration, so we do that inside spec->template, and as you can see above it has a metadata property and spec property in itseldf, so that is like having a conf file inside a conf file. So a pod should have its own configuration inside a deployment configuration file, which is the blueprint of the pod containing information like which port should it open, which image should it use etc.