Betteruptime Edge Deployment

Better Uptime Edge Monitoring

Better Uptime Edge Deployment Multi-region Monitoring CDN Cloudflare Workers Vercel Fly.io Status Page Incident Management Alert

Multi-region Setup

# === Edge Monitoring Configuration ===

from dataclasses import dataclass

@dataclass
class EdgeMonitor:
    region: str
    endpoint: str
    check_locations: str
    expected_latency: str
    alert_threshold: str

monitors = [
    EdgeMonitor("US East (Virginia)",
        "https://us-east.example.com/health",
        "New York, Chicago, Dallas",
        "< 50ms",
        "Response Time > 200ms หรือ Status != 200"),
    EdgeMonitor("US West (Oregon)",
        "https://us-west.example.com/health",
        "Los Angeles, San Francisco, Seattle",
        "< 50ms",
        "Response Time > 200ms หรือ Status != 200"),
    EdgeMonitor("EU West (Frankfurt)",
        "https://eu.example.com/health",
        "London, Frankfurt, Amsterdam",
        "< 80ms",
        "Response Time > 300ms หรือ Status != 200"),
    EdgeMonitor("Asia (Singapore)",
        "https://asia.example.com/health",
        "Singapore, Tokyo, Sydney",
        "< 100ms",
        "Response Time > 400ms หรือ Status != 200"),
    EdgeMonitor("Global (Anycast)",
        "https://example.com/health",
        "All Locations (10+)",
        "< 200ms (worst case)",
        "2+ Locations Fail → Global Alert"),
]

print("=== Edge Monitors ===")
for m in monitors:
    print(f"  [{m.region}]")
    print(f"    URL: {m.endpoint}")
    print(f"    Check from: {m.check_locations}")
    print(f"    Expected: {m.expected_latency}")
    print(f"    Alert: {m.alert_threshold}")

Status Page & Incident

# === Status Page Components ===

@dataclass
class StatusComponent:
    component: str
    group: str
    monitor_id: str
    sla_target: str
    current_status: str

components = [
    StatusComponent("API Gateway",
        "Global Services",
        "HTTP Monitor → api.example.com",
        "99.99% (52.6 min downtime/year)",
        "Operational"),
    StatusComponent("Edge CDN US",
        "US Region",
        "HTTP Monitor → us.example.com",
        "99.95% (4.38 hr downtime/year)",
        "Operational"),
    StatusComponent("Edge CDN EU",
        "EU Region",
        "HTTP Monitor → eu.example.com",
        "99.95%",
        "Operational"),
    StatusComponent("Edge CDN Asia",
        "Asia Region",
        "HTTP Monitor → asia.example.com",
        "99.9% (8.77 hr downtime/year)",
        "Degraded Performance"),
    StatusComponent("Database (Primary)",
        "Infrastructure",
        "TCP Monitor → db-primary:5432",
        "99.99%",
        "Operational"),
    StatusComponent("Background Jobs",
        "Infrastructure",
        "Heartbeat Monitor → /heartbeat",
        "99.9%",
        "Operational"),
]

print("=== Status Page ===")
for c in components:
    emoji = "OK" if c.current_status == "Operational" else "WARN"
    print(f"  [{emoji}] {c.component} ({c.group})")
    print(f"    Monitor: {c.monitor_id}")
    print(f"    SLA: {c.sla_target}")
    print(f"    Status: {c.current_status}")

Incident Workflow

# === Incident Management ===

@dataclass
class IncidentStep:
    step: str
    owner: str
    time_target: str
    action: str
    tool: str

incident_flow = [
    IncidentStep("Alert Triggered",
        "Better Uptime",
        "Instant",
        "ตรวจพบ Monitor Fail ส่ง Alert",
        "Better Uptime → Slack + PagerDuty"),
    IncidentStep("Acknowledge",
        "On-call Engineer",
        "< 5 minutes",
        "กด Acknowledge รับ Incident",
        "PagerDuty App / Slack Button"),
    IncidentStep("Triage",
        "On-call Engineer",
        "< 10 minutes",
        "ตรวจ Scope Regional/Global ตั้ง Priority P1-P4",
        "Better Uptime Dashboard + Grafana"),
    IncidentStep("Update Status Page",
        "On-call Engineer",
        "< 15 minutes",
        "แจ้ง User ว่ากำลังตรวจสอบ",
        "Better Uptime Status Page"),
    IncidentStep("Investigate & Fix",
        "Engineering Team",
        "< 30 min (P1) < 2hr (P2)",
        "หา Root Cause แก้ไข Deploy Fix",
        "Logs Metrics Runbook"),
    IncidentStep("Resolve & Verify",
        "On-call Engineer",
        "หลัง Fix",
        "ตรวจ Monitor กลับ OK Update Status Resolved",
        "Better Uptime + Smoke Test"),
    IncidentStep("Post-mortem",
        "Team Lead",
        "< 48 hours",
        "เขียน Root Cause Timeline Action Items",
        "Confluence / Notion Template"),
]

print("=== Incident Flow ===")
for s in incident_flow:
    print(f"  [{s.step}] Owner: {s.owner}")
    print(f"    Target: {s.time_target}")
    print(f"    Action: {s.action}")
    print(f"    Tool: {s.tool}")

เคล็ดลับ

• Multi-location: ตรวจจากหลาย Location แยก Regional vs Global Issue
• Status Page: สร้าง Status Page แจ้ง User ลดภาระ Support
• Heartbeat: ใช้ Heartbeat ตรวจ Background Job ไม่ใช่แค่ HTTP
• Escalation: ตั้ง Escalation 2+ Location Fail = P1
• Post-mortem: เขียน Post-mortem ทุก Incident ป้องกันเกิดซ้ำ

การนำความรู้ไปประยุกต์ใช้งานจริง

แหล่งเรียนรู้ที่แนะนำ ได้แก่ Official Documentation ที่อัพเดทล่าสุดเสมอ Online Course จาก Coursera Udemy edX ช่อง YouTube คุณภาพทั้งไทยและอังกฤษ และ Community อย่าง Discord Reddit Stack Overflow ที่ช่วยแลกเปลี่ยนประสบการณ์กับนักพัฒนาทั่วโลก

เปรียบเทียบข้อดีและข้อเสีย

จากตารางเปรียบเทียบจะเห็นว่าข้อดีมีมากกว่าข้อเสียอย่างชัดเจน โดยเฉพาะในแง่ของประสิทธิภาพและความสามารถในการ Scale สำหรับข้อเสียส่วนใหญ่สามารถแก้ไขได้ด้วยการเรียนรู้อย่างเป็นระบบและวางแผนทรัพยากรให้เหมาะสม

สรุป

Better Uptime Edge Deployment Multi-region Monitoring Status Page Incident Management Alert Escalation Post-mortem MTTA MTTR Production