Fare Watchtower Operations

Schedule Active

Last Run: 06:00 UTC

3 Alerts Pending

Flights Tracked

847

across 12 routes

Hotels Tracked

412

across 8 cities

Price Drops Today

avg -14.2% from baseline

Est. Savings

$4,280

this week

Uptime

99.7%

last 30 days

Audit Logs

1,204

immutable run records

Flight Price Drops

Route	Base	Now	Drop	Provider	Conf	Alert
SFO→NRT	$1,240	$879	-29.1%	Amadeus	High	Sent
LAX→LHR	$890	$712	-20.0%	Skyscanner	High	Sent
JFK→CDG	$760	$648	-14.7%	Kiwi	Med	Sent
ORD→HND	$1,450	$1,290	-11.0%	Amadeus	Med	Pending

Hotel Price Drops

Hotel	City	Base	Now	Drop	Conf	Alert
Park Hyatt	Tokyo	$520	$385	-26.0%	High	Sent
The Savoy	London	$480	$392	-18.3%	High	Sent
Le Bristol	Paris	$610	$538	-11.8%	Med	Pending
Aman Tokyo	Tokyo	$780	$702	-10.0%	Med	Pending

Route Watchlist

SFO → NRT (Jun 15–22)Target: <$900

LAX → LHR (Jul 1–10)Target: <$750

JFK → CDG (Aug 5–12)Target: <$650

ORD → HND (Sep 20–30)Target: <$1,200

Hotel Watchlist

Park Hyatt Tokyo (Jun 15–22)Target: <$400/n

The Savoy London (Jul 1–10)Target: <$380/n

Le Bristol Paris (Aug 5–12)Target: <$500/n

Aman Tokyo (Sep 20–30)Target: <$680/n

Automation Workflow

🕑

Daily Schedule

06:00 UTC cron

✈️

Flight API

Amadeus / Kiwi

🏨

Hotel API

Booking / Agoda

🔄

Normalize

Schema unify

📈

Compare Baseline

Delta calc

💾

Save to Drive

/fare-watchtower/

📨

Alert Summary

Threshold filter

🌐

Page Dashboard

Static deploy

Drive Archive Structure

/fare-watchtower/
  2026-04-28/
    flights.json — raw flight API response (847 records)
    hotels.json — raw hotel API response (412 records)
    drops.json — filtered price drops exceeding threshold
    run-log.json — execution metadata, timing, errors
    dashboard-data.json — pre-computed dashboard payload
  2026-04-27/
    ... (same structure)

Implementation Snippets

Python — Computer Job
# fare_watchtower.py — MuleRun Computer scheduled task
# Example API endpoints — not production URLs
import os, json, hashlib
from datetime import date
from mulerun import Drive, Page

# Keys from env vars — never hardcoded
FLIGHT_KEY = os.environ["FLIGHT_API_KEY"]
HOTEL_KEY  = os.environ["HOTEL_API_KEY"]
DAY = date.today().isoformat()

def fetch_flights():
    # Example: https://api.example.com/v1/flights/search
    resp = requests.get("https://api.example.com/v1/flights/search",
        headers={"Authorization": f"Bearer {FLIGHT_KEY}"},
        params={"routes": WATCHLIST_ROUTES})
    return resp.json()

def hash_traveler(email):
    return hashlib.sha256(email.encode()).hexdigest()[:12]

# Save to Drive with date partition
drive = Drive(folder="/fare-watchtower")
drive.write(f"{DAY}/flights.json", flights)
drive.write(f"{DAY}/drops.json", drops)
drive.write(f"{DAY}/dashboard-data.json", dashboard)
  

Shell — Cron Setup
# MuleRun Computer scheduled task
$ mulerun computer schedule create \
    --name "fare-watchtower" \
    --cron "0 6 * * *" \
    --command "python3 /opt/fare_watchtower.py" \
    --env FLIGHT_API_KEY="$FLIGHT_API_KEY" \
    --env HOTEL_API_KEY="$HOTEL_API_KEY"

# Deploy static dashboard
$ mulerun page deploy \
    --name "fare-watchtower" \
    --source "/opt/output/index.html"
  

Security Practices

🔒

Environment Variable API Keys

All API credentials stored as encrypted env vars on the Computer instance. Never committed to code or stored in Drive files.

📑

No Secrets in Drive

Drive archive contains only normalized pricing data and metadata. No API keys, tokens, or PII in any stored JSON file.

🛡️

Least-Privilege Drive Folder

The /fare-watchtower/ folder uses scoped write-only access for the automation and read-only access for the dashboard deployer.

🔐

Hashed Traveler IDs

All traveler identifiers are SHA-256 hashed before storage. No reversible PII exists in any archived file.

Outcome Metrics

12.4 hrs

Time Saved / Week

97%

Missed Drops Avoided

$4,280

Avg Weekly Savings

1,204

Immutable Audit Logs