Fly.io Machines Tech Conference 2026

Fly.io Machines

Fly.io Machines Tech Conference Edge Computing Global Deployment Firecracker MicroVM Auto-scaling Anycast Scale to Zero Live Streaming WebSocket Real-time

Fly.io Setup

# === Fly.io Deployment ===

# Install CLI
# curl -L https://fly.io/install.sh | sh
# fly auth login

# Create App
# fly launch --name tech-conf-2026
# fly deploy

# Dockerfile
# FROM node:20-alpine
# WORKDIR /app
# COPY package*.json ./
# RUN npm ci --production
# COPY . .
# EXPOSE 3000
# CMD ["node", "server.js"]

# fly.toml — Configuration
# app = "tech-conf-2026"
# primary_region = "sin"  # Singapore
#
# [build]
#   dockerfile = "Dockerfile"
#
# [http_service]
#   internal_port = 3000
#   force_https = true
#   auto_stop_machines = true
#   auto_start_machines = true
#   min_machines_running = 1
#
# [[vm]]
#   cpu_kind = "shared"
#   cpus = 2
#   memory_mb = 512
#
# [processes]
#   app = "node server.js"
#   worker = "node worker.js"

# Multi-region Deployment
# fly regions add sin nrt lax ams
# fly scale count 2 --region sin
# fly scale count 1 --region nrt
# fly scale count 1 --region lax
# fly scale count 1 --region ams

# Fly Postgres — Multi-region Database
# fly postgres create --name conf-db --region sin
# fly postgres attach conf-db
# fly postgres connect -a conf-db

from dataclasses import dataclass

@dataclass
class FlyApp:
    name: str
    regions: int
    machines: int
    cpu: str
    memory_mb: int
    status: str
    requests_hr: int

apps = [
    FlyApp("conf-web", 4, 6, "shared-2x", 512, "Running", 15000),
    FlyApp("conf-api", 4, 8, "shared-2x", 1024, "Running", 25000),
    FlyApp("conf-ws", 3, 4, "shared-2x", 512, "Running", 8000),
    FlyApp("conf-stream", 2, 3, "performance-2x", 2048, "Running", 5000),
    FlyApp("conf-db", 2, 3, "performance-4x", 4096, "Running", 0),
]

print("=== Fly.io Apps ===")
for a in apps:
    print(f"  [{a.status}] {a.name}")
    print(f"    Regions: {a.regions} | Machines: {a.machines}")
    print(f"    CPU: {a.cpu} | RAM: {a.memory_mb}MB | Req/hr: {a.requests_hr:,}")

Conference Platform

# === Tech Conference Architecture ===

# Components:
# 1. Conference Website — Next.js on Fly.io
# 2. API Server — FastAPI/Express on Fly.io
# 3. WebSocket Server — Real-time Chat & Q&A
# 4. Streaming — HLS/DASH Video Delivery
# 5. Database — Fly Postgres Multi-region
# 6. Cache — Fly Redis (Upstash)
# 7. Storage — S3-compatible (Tigris on Fly)

# Auto-scaling for Keynote
# fly autoscale set min=2 max=20
# # Or via Machines API:
# curl -X POST "https://api.machines.dev/v1/apps/conf-api/machines" \
#   -H "Authorization: Bearer $FLY_API_TOKEN" \
#   -d '{
#     "config": {
#       "image": "registry.fly.io/conf-api:latest",
#       "guest": {"cpus": 2, "memory_mb": 1024},
#       "auto_destroy": true
#     },
#     "region": "sin"
#   }'

# WebSocket — Real-time Q&A
# const WebSocket = require('ws');
# const wss = new WebSocket.Server({ port: 8080 });
#
# wss.on('connection', (ws) => {
#   ws.on('message', (data) => {
#     const msg = JSON.parse(data);
#     if (msg.type === 'question') {
#       // Broadcast to all clients in same session
#       wss.clients.forEach(client => {
#         if (client.readyState === WebSocket.OPEN) {
#           client.send(JSON.stringify({
#             type: 'question',
#             user: msg.user,
#             text: msg.text,
#             timestamp: Date.now()
#           }));
#         }
#       });
#     }
#   });
# });

@dataclass
class ConferenceMetric:
    metric: str
    value: str
    target: str
    status: str

metrics = [
    ConferenceMetric("Registered Attendees", "5,200", "5,000", "Exceeded"),
    ConferenceMetric("Live Viewers (Peak)", "3,800", "3,000", "Exceeded"),
    ConferenceMetric("API Latency (p99)", "45ms", "< 100ms", "OK"),
    ConferenceMetric("WebSocket Connections", "2,100", "< 5,000", "OK"),
    ConferenceMetric("Stream Bitrate", "1080p 6Mbps", "720p+", "OK"),
    ConferenceMetric("CDN Cache Hit", "95%", "> 90%", "OK"),
    ConferenceMetric("Error Rate", "0.02%", "< 0.1%", "OK"),
    ConferenceMetric("Monthly Cost", "$280", "< $500", "Under Budget"),
]

print("\nConference Dashboard:")
for m in metrics:
    print(f"  [{m.status}] {m.metric}: {m.value} (target: {m.target})")

Cost และ Optimization

# === Cost Optimization ===

# Fly.io Pricing (approximate)
# shared-cpu-1x, 256MB: $1.94/mo (running 24/7)
# shared-cpu-2x, 512MB: $3.88/mo
# performance-2x, 2GB: $31.00/mo
# Scale to Zero: pay only when running

# Conference Cost Breakdown
cost_items = {
    "Web App (4 regions, 6 machines)": "$23/mo",
    "API Server (4 regions, 8 machines)": "$62/mo",
    "WebSocket (3 regions, 4 machines)": "$16/mo",
    "Streaming (2 regions, 3 machines)": "$93/mo",
    "Postgres (2 regions, HA)": "$50/mo",
    "Bandwidth (500GB)": "$25/mo",
    "Persistent Volumes": "$15/mo",
    "Total": "$284/month",
    "During Conference (scaled up)": "~$50/day extra",
}

print("Cost Breakdown:")
for item, cost in cost_items.items():
    print(f"  {item}: {cost}")

# Scale to Zero Strategy
# Before conference: min_machines = 1
# During keynote: auto-scale to 20
# After conference: scale to zero
# Cost saving: ~70% vs always-on

optimization_tips = [
    "Scale to Zero: ใช้ auto_stop เมื่อไม่มี Traffic",
    "Shared CPU: ใช้ Shared CPU สำหรับ Web และ API",
    "Performance: ใช้ Performance CPU สำหรับ Streaming เท่านั้น",
    "Region: Deploy เฉพาะ Region ที่มีผู้เข้าร่วม",
    "CDN: ใช้ Fly CDN Cache Static Assets",
    "Pre-warm: Scale up ก่อน Keynote 15 นาที",
    "Monitor: ใช้ fly logs และ fly status ติดตาม Real-time",
]

print(f"\n\nOptimization Tips:")
for i, t in enumerate(optimization_tips, 1):
    print(f"  {i}. {t}")

เคล็ดลับ

• Scale to Zero: auto_stop_machines ประหยัดเงินมาก
• Multi-region: Deploy ใกล้ผู้ใช้ Latency ต่ำ
• Postgres: ใช้ Fly Postgres Multi-region HA
• Pre-warm: Scale up ก่อนงาน Event ป้องกัน Cold Start
• CLI: fly deploy ง่ายสุด ใช้ได้ทุก Language

Fly.io คืออะไร

Platform Deploy ใกล้ผู้ใช้ Firecracker MicroVM 30+ Region Docker Auto-scaling Anycast Persistent Volume Free Tier CLI ง่าย

สรุป

Fly.io Machines Tech Conference Edge Computing Global Deployment Firecracker MicroVM Auto-scaling Scale to Zero WebSocket Streaming Multi-region Postgres CDN Production