loadbalancer

A generic loadbalancer built on GoLang to help learn go and its concurrency. To register a host, it needs a list of upstream servers, hostname, port, healthcheck and routing algorithm to be used for that host. Multiple hosts can be registered at a time. The healthchecks dynamically update the available collection of upstream servers.

Setup

Build and Run:

chmod +x build.sh
./build

Run:

./bin/loadbalancer

Test: use the docker-compose file to bring up example servers. You can change services and edit config.yaml to your requirement

docker-compose -f ./examples/docker-compose.yaml up -d 

Features

This load-balancer features three routing algorithms:

1. Random
2. Round Robin
3. Weighted Round Robin

Configurations

The loadbalancer will read yaml configurations from the root directory. You can specify filename in main package when you build. Default is config.yaml

Attributes	Values	type
env	dev, prod	string
logLevel	info, debug	string
host.name	FQDN for host	string
host.scheme	random, roundrobin, weightedroundrobin	string
host.timeout	Request timeout duration in seconds	int
host.health	healthcheck for upstream	string
host.interval	healthcheck interval	int
servers.name	FQDN for upstream	string
servers.weight	weight for weightedRR	int

A sample config file

env: dev
logLevel: info
hosts: 
- name: localhost 
  port: 3000
  scheme: roundrobin
  health: /health
  interval: 10
  timeout: 8
  servers:
  - name: localhost:9081
    weight: 5
    health: 
  - name: localhost:9082
    weight: 3
  - name: localhost:9083
    weight: 2
- name: localhost 
  port: 3001
  scheme: random
  health: /health
  interval: 10
  timeout: 8
  servers:
  - name: localhost:9081
    weight: 5
    health: 
  - name: localhost:9082
    weight: 3
  - name: localhost:9083
    weight: 2

Stress Test

Routing: Weighted Round Robin

Summary:

Requests: 100000
Total: 101.7378 secs
Slowest: 0.3187 secs
Fastest: 0.0006 secs
Average: 0.0496 secs
Requests/sec: 982.9189

Total data: 17900000 bytes
Size/request: 179 bytes

Status code distribution:

[200] 100000 responses

Latency distribution:

10% in 0.0031 secs
25% in 0.0428 secs
50% in 0.0478 secs
75% in 0.0645 secs
90% in 0.0857 secs
95% in 0.1009 secs
99% in 0.1380 secs

Errors:

0

Progress

• Setup Project

• Create initializer Sequence

• Implement proxy

• Implement random, rr and weighted rr routing

• [Added Late] Refactor packages without side effects

• Add dynamic routing with healthchecks

• Multiple concurrent hosts

• Unit tests

• Graceful exit

• Stress test and attach result

• Add benchmarks

• Proxy Timeouts