How To Configure A High-availability Load Balancing Environment On Alibaba Cloud Singapore Cn2

2026-04-09 10:29:25
Current Location： Blog > Singapore CN2

1.

overview and objectives

1) goal: build a high-availability load balancing in alibaba cloud singapore cn2 environment to ensure 99.99% business availability.
2) applicable scenarios: multinational e-commerce, api gateways, game login servers and other delay-sensitive services.
3) network characteristics: cn2 is directly connected back to china, with an average one-way delay of about 40ms (singapore -> eastern china).
4) core components: ecs (backend), slb/haproxy (load balancing), alibaba cloud cdn, anti-ddos.
5) sla indicators: failover <10 seconds, health check interval 5s, timeout 2s.

2.

architecture design

1) dual availability zone deployment: place at least one set of ecs in singapore-a and singapore-b.
2) cross-machine room load: use alibaba cloud slb (off-site redundancy) + private network connection for heartbeat.
3) session retention: select the source ip or cookie stickiness according to the business to avoid session loss.
4) health check: http get /health, interval 5s, failure threshold 3 times and recovery threshold 2 times.
5) monitoring alarms: combined with cloud monitoring, set cpu>70%, delay>200ms alarms and automatically expand the capacity.

3.

instance and network configuration examples

1) sample backend server specifications: ecs.c6.large, 2 vcpu, 4gb memory, public network bandwidth 100mbps.
2) operating system: ubuntu 20.04 lts, kernel optimization tcp_tw_reuse=1, net.ipv4.tcp_fin_timeout=30.
3) connection number optimization: ulimit -n 200000, nginx worker_connections 4096.
4) anti-ddos: turn on anti-ddos pro basic, threshold 1gbps, burst traffic protection for 5 minutes.
5) example response indicators: a single ecs handles static requests at 8000 rps, with an average delay of 18ms.
node ip role weight
ecs-1 10.0.1.11 backendweb 120
ecs-2 10.0.2.12 backendweb 120
haproxy 10.0.0.5 front-end load -

4.

haproxy configuration and examples

1) front-end monitoring: bind 0.0.0.0:80 and use option httpchk to set the health check path.
2) back-end strategy: use roundrobin or weight-based load distribution.
3) timeout settings: timeout connect 5s, timeout client 30s, timeout server 30s.
4) session persistence: stick on src or cookie, depending on the business.
5) example snippet (brief display):
 frontend http_front
  bind *:80
  default_backend web_back

backend web_back
  mode http
  balance roundrobin
  option httpchk get /health
  server web1 10.0.1.11:80 weight 120 check
  server web2 10.0.2.12:80 weight 120 check

5.

real cases and fault recovery

1) case company: a cross-border e-commerce company deploys international settlement services in singapore cn2.
2) configuration: 2 az, 3 ecs.c6.large each, front-end slb + two haproxy for active-active.
3) traffic peak: single-day peak value 350k rps, peak bandwidth 420mbps, 86% request mitigation by cdn cache.
4) fault drill: shut down the singapore-a host group, the failover time is 6 seconds on average, and no business interruption is reported.
5) optimization effect: using the cn2 backend link, the average delay is 15ms lower than that of the ordinary public network, and the user experience is significantly improved.

6.

operation, maintenance and security suggestions

1) cdn offloading: static resources are cached by alibaba cloud cdn, reducing the pressure and cost of returning to the source.
2) ddos protection: enable anti-ddos pro and combine it with the whitelist rate limiting strategy.
3) automatic expansion: combined with alibaba cloud elastic scaling, automatically add/delete ecs backends based on tps.
4) logs and auditing: centrally collect access logs, set abnormal traffic threshold alarms and trigger script flow limiting.
5) regular drills: conduct disaster recovery drills every quarter to verify the health check and automatic switching process.

singapore cn2
Tags：alibaba cloud singapore cn2 load balancing ha high availability ecscdnddos protection More»
Previous article： Research On The Performance Of Singapore’s Cn2 Lines In Corporate Telecommuting And Video Conferencing
Latest articles
How Do Geographical Restrictions Caused By Non-japanese Native Ip Affect Shopping, Streaming And Payment Experiences?
Practical Experience Sharing On The Security And Compliance Requirements Of Singapore Servers
Singapore Cmi Vps Control Panel Operation Tutorial And Common Function Configuration Guide
Which Industries Are Google Cloud Korea Servers Suitable For And Analysis Of Typical Deployment Cases?
Taiwan Vps Stable Deployment Practical Experience Sharing And Common Troubleshooting
Follow Compliance Requirements And Safely Use Vietnamese Native Residential Ip To Avoid The Risk Of Account Ban
From The Perspective Of Latency And Link Stability, Why Korean Servers Are Better At Carrying Cross-border Traffic?
Japan, Hong Kong And The United States Vps Comparison Case Measured Access Speed Differences In Different Regions
How To Use Your Budget To Decide The Best Time To Buy In The Us High Defense Server Rankings
From The Network Operator's Perspective, What Should I Do If Taiwan's Server Is Stuck? How To Communicate With Isp To Optimize Link Quality?
Popular tags
Video Live Broadcast Optimization Link Diagnosis Simulator Smooth Experience Enterprise-level Network Architecture CN2 Service Budget Optimization Server Advantages Backup And Disaster Recovery Overloaded Ip Geographical Location Vultr Miui Operator Prompt Copywriting Access Automatic Expansion And Contraction Customer Base Network Performance Hosting Cloud Migration Cases Amazon Operates Japan Station Language Optimization Tariff Performance Comparison Wechat Internet Marketing Data Privacy Preventing Theft Flextronics
Related Articles