sre.md

Site Reliability Engineer Challenge

Introduction

In Lalamove we are scaling rapidly and we need SREs to help us prevent downtime and provide visibility for our engineers into what things may go wrong.

Requirement

1. We value a clean, simple working solution.
2. The application must be run in Docker, candidate must provide docker-compose.yml and start.sh bash script at the root of the project, which should setup all relevant dependencies.
3. The solution must be written in Golang.
4. Candidate must submit the project as a private git repository (github.com / bitbucket.com / gitlab.com). Repository must grant read access with the public SSH key.
5. Having unit/integration tests is a strong bonus.
6. The solution must be production ready with appropriate documentation.

Problem Statement

You will build a configurable http ingress proxy with built in observability that can be used as a kubernetes sidecar.

Feature requirements:

• Support prometheus metrics (design appropriate metrics to be exposed).
• Support opentracing. Should either propagate the tracing span from the caller or inject a new trace.

Bonus features:

• Circuit breaker implementation for request failures.
• Configurable rate limiting.

Example yaml config file that will be used as an input:

port: 80
metricsEndpoint: "/metrics"
metricsPort: 9100

tracing:
    samplingServer: "jaeger:6831"

paths: 
- path: "/foo*"
  method: "GET"
  forwardPort: 8080
  # max number of times the proxy can retry a failed request (5xx status code)
  maxRetry: 3
  circuitBreaking:
    # number of consecutive errors after which the circuit breaker should reject incoming requests 
    maxError: 3
    # time for which incoming requests should be rejected
    rejectionPeriod: 5s
  rateLimiting:
    # maximum number of requests to be allowed in the given time interval
    maxAmount: 5000
    expiration: 5s

Questions? We love to answer: [email protected]