Access data produced in a pod from a different pod in Kubernetes - kubernetes

Question

I have a pod, called pod1 that runs a function which produces an array with random unordered integer elements and have also a service for this pod.
In addition, in the same cluster i have an other pod, called pod2 which runs a sorting function.
The pod1 has the following python code:
from flask import Flask,jsonify
app = Flask(__name__)
numbers= [1, 3, 5, 7, 9, 10, 2, 8]
#app.route('/numbers')
def get_numbers():
return jsonify(incomes)
if __name__=="__main__":
app.run(port='8080', debug=True)
Also the deployment file of the above pod is the following:
apiVersion: apps/v1
kind: Deployment
metadata:
name: flaskapp
labels:
app: flaskapp
spec:
selector:
matchLabels:
app: flaskapp
replicas: 1
template:
metadata:
labels:
app: flaskapp
spec:
containers:
- name: flaskapp
image: application:latest
imagePullPolicy: IfNotPresent
ports:
- containerPort: 8080
---
apiVersion: v1
kind: Service
metadata:
name: flaskapp
spec:
ports:
- name: http
port: 8080
targetPort: 8080
selector:
app: flaskapp
In addition i have also another pod whi acts like a client and has the folloowing code:
import requests
x = requests.get('http://flaskapp:8080/numbers')
print(x.text)
in the http the flaskapp is the name of the service of pod1.
So when i apply the deployment file of pod2, the client pod is running succesfully but when i use the kubectl logs for that pod in order to see the expected output, nothing happens.
Is the definition of the requests.get('http://flaskapp:8080/numbers') correct? What i am doing wrong?

Answer 1

It would help by using the correct terms in your question. A pod is simply an instance of a service and there are usually more than 1 related to the same service. When you are talking about 2 pods, it sounds like you are talking about two separate services, for example Processer and RandomNumberGenerator or something.
In microservices, services should only talk to each other via their published ports, this restricts implementation details to the service and works correctly with the way that cluster orhcestrators work (e.g. ephermeral pods, erroring pods etc.).
In your case, as #botje already suggested, if the request is synchronous, the second service can ask the first service for the array of numbers explicitly (send a request to the service using internal DNS) and process the results. If it is asynchronous, then you could use something like a queue and put the numbers into it, to have the processor take the list of numbers from the queue and process it. You would have to build the queue, it is not part of Kubernetes.

Answer 2

Assuming you have:
Pod generator, which exposes a Flask app on port 8080
Service generatorSvc, which targets the generator Pod and exposes port 8080 as well
Pod client,
then in client you can use
r = requests.get('http://generatorSvc:8080/numbers')
EDIT Since your service has port: 8000, you should connect to http://generatorSvc:8000/numbers instead.