Welcome to this DevOps/SRE technical test at AirFund.
The goal is to assess your ability to monitor, diagnose, and fix a problem using Kubernetes only.

This stack includes 2 services:

• api-service (Go): an HTTP API that sends requests to the worker-service.
• worker-service (Python): performs the processing.
• A bash script generate_request.sh that sends a series of requests to api-service to test the stability and robustness of the stack under load.

flowchart TD
    A[Bash Script] -->|GET /process| B[api-service]
    B -->|POST /do-work| C[worker-service]

It is deployed in a Kubernetes cluster with a functional monitoring stack (Grafana, Loki, Prometheus).

👉 In the current state, we are unable to achieve 100% success for a series of 100 requests executed with this script.

Maximum duration: 1h

1. Create a Grafana Dashboard to monitor the stack
2. Find and diagnose the problem using this dashboard
3. Propose and apply a fix using Kubernetes only (no code modification) to achieve 100% success for a series of 100 requests executed with this script.

• Docker
• kubectl
• kind or minikube (recommended: kind)
• skaffold

1. Create a local cluster

   kind create cluster --name airfund-test

2. Deploy the stack

   skaffold dev

3. Once deployment is complete, the services are available on the following ports:

   • api-service: http://localhost:8080
   • worker-service: http://localhost:8081
   • grafana: http://localhost:8082

   The Grafana login credentials are:

   • username: admin
   • password: devops@airfund2025

4. Run the load test

   bash generate_request.sh 100

• Do not modify the code of the services
• Explain your approach
• Internet and AI are authorized to search, compare or ask for advice
• If blocked, ask questions or ask for help

Good luck — show us how you observe, analyze and stabilize a system in real conditions 🚀

— AirFund Team