This a test project for interview purposes. It partially contains data of a different project described below in the details.
I used the image of a Spring app with PostgreSQL connection from Dockerhub:
https://hub.docker.com/r/deedat95/docker-spring-boot-postgres
The app code and API description is available on Github, from the same author I assume:
https://github.com/Nibraz15/backend-interview-task
The necessary containers and network can be started by docker-compose from the root of this repository as follows:
docker-compose -f .\mpl-docker-compose.yaml start
You should see containers running in Docker, with he port 8080 exposed to localhost:
It is possible to test the API with Postman on localhost:8080 host.
The testing is easier with importing the test collection located in /postman directory of this repository.
All the tests should pass successfully if everything runs well in the Docker:
The fastest way was to use Kompose and transcript my docker-compose yaml to k8s resources via kompose convert -f mpl-docker-compose.yaml.
It was necessary to crerate db-service.yaml manually, as it had no port defined in docker-compose yaml.
Created resources can be found at /k8s/app directory of this repository.
After (optional) creating of a new namespace:
kubectl create namespace mpl-app
it is possible to deploy all the resources from the /k8s/app directory:
kubectl apply -f .\ -n mpl-app
After a while, there should be 2 pods running:
For the test of this simple deployment, it is necesary to expose the service endpoint of the pod to the localhost.
After kubectl get pods -n mpl-app we know which pod's to forward - ideally to the local port different from 8080 where our Docker instance runs:
kubectl port-forward -n mpl-app app-658d574cc5-g9jq4 8081:8080
In /postman directory of this repository there is a powerplantTest_k8s.postman_collection.json file prepared to import to Postman and test on localhost:8081.
The result should look like this:
The easiest way to add system monitoring to the cluster is using a helm chart of a whole stack from prometheus-community helm repository.
After installing it ...
... and getting necessary port and pod name ...
... it is possible to forward the port to localhost to access Gafana UI:
kubectl port-forward -n prometheus prometheus-grafana-7b988bb846-xv9gg 3000:3000
Default username 'admin' and password 'prom-operator' can be found in Base64 in prometehus-grafana secret:
After logging in on localhost:3000 address, we can choose some already pre-created dashboards in UI:
