OpsfleetAgent Bootcamp

Before the workshop

Get ready for day 1

About 20 minutes, a few days ahead: install three tools, sign in to AWS and run one check. Whatever fails here would otherwise cost you the first hour of the workshop.

  1. Install the tools

  2. Clone and configure

    terminal
    git clone https://github.com/Opsfleet/genai-launchpad.git datastream-bootcamp
cd datastream-bootcamp
uv sync
    terminal · macOS / Linux
    cp .env.example .env
    terminal · Windows PowerShell
    copy .env.example .env

    In .env set PARTICIPANT to the name your instructor gave you; AWS_PROFILE=bootcamp is already set.

  3. Sign in to AWS

    Open the sign-in URL from your invitation email, sign in with the username and temporary password, and choose your own password. Then create a key for the command line: account menu (top right) > Security credentials > Create access key > Command Line Interface (CLI), and save it as the bootcamp profile:

    terminal
    aws configure --profile bootcamp   # key, secret, region from the email, output json

    To look around in the console (your agents, logs, traces), use the switch role link from the email each time you sign in. Lost your password? Ask your instructor for a new invitation.

  4. Run the pre-flight check

    terminal
    uv run bootcamp.py doctor

    It creates nothing, costs a fraction of a cent (one tiny model call) and finishes in seconds. Every line should start with ✔; send the output to your instructor either way.

    terminal · expected outputread only
    Checking this laptop for the bootcamp (up to ~30 s; nothing is created)...
✔ Operating system: macOS 15.5 (arm64)
✔ Python: 3.13.13
✔ uv: uv 0.11.8
✔ Terraform: Terraform v1.15.0 (>= 1.9)
✔ AWS CLI: aws-cli/2.34.40
✔ Disk space: 163.7 GiB free
✔ AWS credentials: arn:aws:iam::123456789012:user/bootcamp/bootcamp-alice (from bootcamp)
✔ .env PARTICIPANT: alice (.env), region us-east-1
✔ Participant role: assumed bootcamp-participant-alice in us-east-1
✔ Gateway settings: https://llm.bootcamp.example.com
✔ LiteLLM call: gpt-6-luna answered in 1.0 s
✔ LiteLLM budget: $5.0000 left of $5
✔ Bedrock: direct Bedrock is denied, as designed (models go through LiteLLM)
✔ Network: Cognito: https://cognito-idp.us-east-1.amazonaws.com/ answered HTTP 400
✔ Network: National Weather Service: https://api.weather.gov/ answered HTTP 200
✔ Network: GitHub: https://github.com/ answered HTTP 200

16 pass. Ready for day 1.

When a line fails

CheckUsually meansFix
Terraform, AWS CLI, uvNot installed or too oldInstall from the links above, open a new terminal
AWS credentialsYour access key is wrong or deleted, or AWS_PROFILE is wrongaws configure --profile bootcamp with a key from the console (Security credentials)
.env PARTICIPANTBlank or malformed nameSet the name your instructor gave you in .env
Participant roleYour name is not on the deployed roster, or the key belongs to another userCheck the spelling, then tell your instructor
Gateway settingsWrong region, or your key was not synced yetCheck BOOTCAMP_REGION, then tell your instructor
LiteLLM call / budgetGateway unreachable (HTTP 503 outside 08:00-22:00 Israel time, when it is off), or your budget is used upRetry within the gateway's hours; otherwise check VPN/proxy, then tell your instructor
BedrockWARN only: direct Bedrock should be deniedTell your instructor (nothing to fix on your side)
NetworkA proxy or firewall blocks an HTTPS hostAsk IT to allow the hosts listed below

Corporate networks

Behind a proxy or firewall, HTTPS must reach:

  • Cognito: https://cognito-idp.<region>.amazonaws.com/
  • National Weather Service: https://api.weather.gov/
  • GitHub: https://github.com/
  • AWS APIs (*.amazonaws.com), the bootcamp gateway and site your instructor shares
  • Python packages and Terraform providers (pypi.org, files.pythonhosted.org, registry.terraform.io, releases.hashicorp.com)

Stuck? Run uv run bootcamp.py doctor --verbose and send the output to your instructor before the workshop.