DevOps Interview Environment

Quick commands via Makefile:

make setup              # First-time setup
make grant-ssm          # Generate SSM credentials for candidate
make check-bootstrap    # Check if instance is ready (~5 min)
make refresh-instance   # Fresh instance between candidates
make grant-ssh SSH_KEY='...' CANDIDATE_IP=x.x.x.x  # Optional SSH access
make revoke-ssh CANDIDATE_IP=x.x.x.x               # Revoke SSH access

First-time Setup

make setup
# Wait ~5 min, then check bootstrap status:
make check-bootstrap

Share SSM access with Candidate

make grant-ssm
# Share output with candidate; creds expire in 1 hour(!)

SSH access (Optional)

SSH is configured but blocked by default. To enable:

# Grant access
make grant-ssh \
  SSH_KEY='ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIJ...' \
  CANDIDATE_IP=203.0.113.10

# Share with candidate: ssh test@<IP>
terraform output -raw instance_public_ip

# Revoke when done
make revoke-ssh CANDIDATE_IP=203.0.113.10

Between Candidates

make refresh-instance
# Creates fresh instance in ~5 min
# Old credentials auto-expire

Challenges

The candidate is given a broken Kubernetes setup (templates/all-challenges.yaml) with three independent failures to diagnose and fix. The goal is to get the LLM service reachable at http://llm.interview.local.

Solutions (spoilers) are in SOLUTIONS.md.