Getting null pointer exception after deployment in GCP

Question

I have a .NET application divided in microservices. One of the microservices is the orchestrator and authenticaticator: this microservice communicates to 3 other microservices and the intercommunication is done via REST. The code implemented is fully functional locally and each microservice was also containarized with docker and tested locally (also bug free). The problems arise when we try to test our services in GCP: the error is always the same, being it a null pointer exception, as can be seen in the image below: We have defined environment variables in the code that match with the .yml file which is used to make the deployment. This is the .yml that we are using:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: orchauth
  namespace: default
  labels:
    app: orchauth
spec:
  strategy:
    type: Recreate
  replicas: 2
  selector:
    matchLabels:
      app: orchauth
  template:
    metadata:
      labels:
        app: orchauth
    spec:
      serviceAccountName: default
      containers:
      - name: orchauth
        image: "eu.gcr.io/boomoseries-api/orchmicroservice"
        ports:
        - containerPort: 80
        env:
        - name: SEARCH_HOST
          value: "http://search.default"
        - name: USERS_HOST
          value: "http://users.default"
        - name: PREFS_HOST
          value: "http://prefs.default"
---
apiVersion: v1
kind: Service
metadata:
  name: orchauth
spec:
  selector:
    app: orchauth
  ports:
  - name: http
    protocol: TCP
    port: 80
    targetPort: 80

The other microservices are implemented similarly in the file. We also defined the microservices environment variables in the code, correctly:

private static readonly string microservicesBaseURL = Environment.GetEnvironmentVariable("SEARCH_HOST");

private static readonly string microservicesBaseURL = Environment.GetEnvironmentVariable("USERS_HOST");

private static readonly string microservicesBaseURL = Environment.GetEnvironmentVariable("PREFS_HOST");

This is the ingress .yml file:

apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: boomoseries-ingress
spec:
  ingressClassName: nginx
  defaultBackend:
    service:
      name: orchauth
      port:
        number: 80

The ingress communicates with the microservice OrchAuth, which is the one giving the error.

The error occurs in the request: from my understanding, the request variable is null and the null pointer exception occurs here. But I can't understand why, since locally and containarized the service works with no issues. The code is presented below (the class SearchRESTComunicationServiceBooks):

public async Task<List<BookDTO>> ObtainBooksByRating(string type, double minRating)
        {
            List<BookDTO> booksDtos = new();
            //Makes the requests to different microservices
            Stopwatch stopwatch = new Stopwatch();
            stopwatch.Start();
            var request = httpClient.GetAsync(microservicesBaseURL + "?type=" + type + "&minRating=" + minRating);

            stopwatch.Stop();
            System.Diagnostics.Debug.WriteLine(stopwatch.ElapsedMilliseconds);
            //Get the responses
            var response = request.Result;

            var responseString = await response.Content.ReadAsStringAsync();
            List<BookDTO> deserializedBook = JsonConvert.DeserializeObject<List<BookDTO>>(responseString);
            foreach (var item in deserializedBook)
            {
                booksDtos.Add(item);
            }
            return booksDtos;

Note: I am fully aware that the code has bad implementation practices, but that is not the main concern. Thank you for the time reading.

Answer 1

So the problem was the fact that in my .yml file I was only configuring my services to use the port 80 (http). In my startup, I was using Hsts, meaning that there would always be a redirection to the https port (which was configured to 443). After adding this port to the .yml file, the problem was solved.