CEX/DEX Latency measurement with centralization

Vultr Multi-Region Latency : Orchestration > Deploy > Tester > Aggregator

Measure HTTP latency to popular CEX/DEX endpoints while provisioning short‑lived Vultr instances across multiple regions. The script runs for a selected duration (0 = single pass from each region, 1, 5, 15, or 60 minutes), aggregates results, estimates costs, and prompts to destroy instances (auto‑destroys after 30s if no answer) to avoid unnecessary charges.

Features

• Multi‑region instance provisioning on Vultr (Cloud Compute vc2-1c-2gb, Ubuntu 22.04).
• Latency measurements to curated CEX/DEX endpoints using async HTTP (aiohttp).
• Interactive duration selector: 0 (single pass), 1, 5, 15, or 60 minutes (1h also accepted).
• Aggregated results (pivot + Top 10) printed to console and saved to CSV.
• Cost estimation proportional to the selected test duration.
• Safe teardown: prompts for destruction and defaults to destroy after 30s of inactivity.
• Structured logging to console and rotating file latency-multi-geo.log.
• Colored output for average latencies: < 75 ms (green), 75–200 ms (orange), > 200 ms (red).

Prerequisites

• Python 3.10+
• A Vultr API key with permissions to create/destroy instances
• macOS/Linux shell (tested on macOS)

Installation / Setup

1. Create and activate a virtual environment

python3 -m venv .venv
source .venv/bin/activate

2. Install dependencies

pip install -r requirements.txt

3. Configure environment variables

• Create a .env file (already git‑ignored) and set your key:

# .env
VULTR_API_KEY=your-vultr-api-key

• Or export it in your shell:

export VULTR_API_KEY=your-vultr-api-key

SSH key setup (recommended)

To connect to instances without passwords, attach an SSH key at creation time and use your local private key for ssh/scp.

1. Generate (if needed) an Ed25519 key pair

ssh-keygen -t ed25519 -C "you@example.com" -f ~/.ssh/id_ed25519 -N ""
eval "$(ssh-agent -s)" && ssh-add ~/.ssh/id_ed25519

Example (your values will differ):

ssh-keygen -t ed25519 -C "XXXXXXX@gmail.com"
# ... choose ~/.ssh/id_ed25519 (or accept default), set passphrase if desired
cat ~/.ssh/id_ed25519.pub
# -> ssh-ed25519 AAAAC3Nz... petirico@gmail.com  (single line)

2. Add the PUBLIC key to Vultr (Account → SSH Keys), then fetch its ID

export VULTR_API_KEY=your-vultr-api-key
curl -s -H "Authorization: Bearer $VULTR_API_KEY" https://api.vultr.com/v2/ssh-keys \
| python -c 'import sys,json; d=json.load(sys.stdin); [print(k["id"], k["name"]) for k in d.get("ssh_keys", [])]'

Example output:

{
  "id": "044d0839-64ae-443b-836a-017fe1776937",
  "name": "ssh-ed25519 "
}

3. Configure .env for the script

# Attach the Vultr key ID at instance creation (array supported via comma‑separated list)
VULTR_SSH_KEY_IDS=your-ssh-key-id

# Local private key used by ssh/scp
SSH_KEY_PATH=/Users/you/.ssh/id_ed25519

# Optional fallback: inject a public key via cloud‑init user_data
# (used only if needed; API attachment is preferred)
# SSH_PUBLIC_KEY_PATH=/Users/you/.ssh/id_ed25519.pub
# or
# SSH_PUBLIC_KEY="ssh-ed25519 AAAA... you@example.com"

# Optional SSH waiting/timeout tuning
# SSH_CONNECT_TIMEOUT=30
# SSH_WAIT_RETRIES=20
# SSH_WAIT_DELAY=10

4. Verify the key matches

ssh-keygen -lf ~/.ssh/id_ed25519.pub
ssh -o StrictHostKeyChecking=no -i ~/.ssh/id_ed25519 root@<instance_ip>

If the last command connects as root without asking for a password, your setup is correct and fully programmatic (no interactive prompts).

Usage

Run the main script:

source .venv/bin/activate
python latency-multi-geo.py

You will be prompted to choose a duration: 0 (single pass), 1, 5, 15, or 60 minutes (1h also accepted). The script will:

• Create instances in default regions: Tokyo (nrt), Singapore (sgp), Frankfurt (fra), New York (ewr), Seoul (icn).
• Wait for instances to become active.
• Perform repeated latency measurements during the selected time window.
• Print a pivot table and Top 10 latencies.
• Save results to a timestamped CSV file: vultr_latency_test_YYYYMMDD_HHMMSS_<duration>m.csv.
• Estimate the test cost and ask whether to destroy instances.
  • If you do not answer within 30 seconds, instances are destroyed automatically.

Output

• CSV: vultr_latency_test_<timestamp>_<duration>m.csv
• Log file: latency-multi-geo.log (rotating)
• Console summary: pivot table (average latency per region) and Top‑10 best latencies

Preview

SSH key saved in Vultr UI:

Billing Notes (Vultr)

• Instances are billed hourly/minute with a monthly cap. There is no long‑term commitment.
• Stopped instances still incur charges; only destroying them stops billing.
• This project prompts for teardown and destroys instances automatically after 30s if unattended.

Configuration

• Regions: update regions_to_deploy in latency-multi-geo.py (default: nrt, sgp, fra, ewr, icn).
• Plan/OS: VULTR_PLAN_ID = "vc2-1c-2gb", VULTR_OS_ID = 1743 (Ubuntu 22.04).
• Endpoints: see REGION_EXCHANGE_MAP inside latency-multi-geo.py.
• Measurement interval: currently 30 seconds between iterations.

Troubleshooting

• 400 Invalid user_data (check base64 encoding): The script encodes user_data in Base64 as required. If it persists, verify your API key and account permissions.
• Permission denied (publickey,password): Ensure your Vultr SSH key is attached (VULTR_SSH_KEY_IDS) and your local private key path is set (SSH_KEY_PATH). Recreate instances after fixing keys.
• SSH not ready / timeouts: Increase SSH_WAIT_RETRIES/SSH_WAIT_DELAY or SSH_CONNECT_TIMEOUT. Ensure no firewall blocks port 22.
• Permission errors on API: Ensure the API key is active and authorized for instance create/delete.
• No results saved: If the measurement window completes without successful requests, the script will print a warning and skip CSV output.

Certificates / Verification

• HTTPS calls use requests/aiohttp default certificate validation. If you need custom CAs or to disable verification temporarily, adapt the client session accordingly.
• SSH host keys are accepted with StrictHostKeyChecking=no for automation; tighten this for production by managing known_hosts.

Security

• Secrets are stored in .env which is listed in .gitignore.
• Never commit API keys to version control.

License

This project is provided as‑is for demonstration and testing purposes. Review and adapt before production use.