负载均衡与健康检查

负载均衡

你可以将多个相同类型的代理加入到同一个 group 中，以实现负载均衡的能力。

目前支持的代理类型包括：tcp, http, tcpmux。

# frpc.toml
    [[proxies]]
    name = "test1"
    type = "tcp"
    localPort = 8080
    remotePort = 80
    loadBalancer.group = "web"
    loadBalancer.groupKey = "123"
    
    [[proxies]]
    name = "test2"
    type = "tcp"
    localPort = 8081
    remotePort = 80
    loadBalancer.group = "web"
    loadBalancer.groupKey = "123"

当用户连接 frps 服务器的 80 端口时，frps 会将接收到的用户连接随机分发给其中一个存活的代理。这可以确保即使一台 frpc 机器挂掉，仍然有其他节点能够提供服务。

对于 tcp 类型代理，需要确保 groupKey 相同以进行权限验证，同时 remotePort 也需一致。

对于 http 类型代理，需要保证 groupKey, customDomains(自定义域名)，subdomain 和 locations 相同。

健康检查

通过给代理配置健康检查参数，可以在要反向代理的服务出现故障时，将该服务从 frps 中摘除。结合负载均衡的功能，这可用于实现高可用架构，避免服务单点故障。

要启用健康检查功能，需要在每个代理的配置中添加 healthCheck.type = {type}。

目前支持的类型有 tcp 和 http。

对于 tcp，只要能够建立连接，即认为服务正常。
对于 http，会发送一个 HTTP 请求，服务需要返回状态码 2xx 才会被视为正常。

以下是 tcp 示例配置：

[[proxies]]
    name = "test1"
    type = "tcp"
    localPort = 22
    remotePort = 6000
    # 启用健康检查，类型为 tcp
    healthCheck.type = "tcp"
    # 建立连接超时时间为 3 秒
    healthCheck.timeoutSeconds = 3
    # 连续 3 次检查失败，此 proxy 会被摘除
    healthCheck.maxFailed = 3
    # 每隔 10 秒进行一次健康检查
    healthCheck.intervalSeconds = 10

以下是 http 示例配置：

[[proxies]]
    name = "web"
    type = "http"
    localIP = "127.0.0.1"
    localPort = 80
    customDomains = ["test.yourdomain.com"]
    # 启用健康检查，类型为 http
    healthCheck.type = "http"
    # 健康检查发送 http 请求的 path，后端服务需要返回 2xx 的 http 状态码
    healthCheck.path = "/status"
    healthCheck.timeoutSeconds = 3
    healthCheck.maxFailed = 3
    healthCheck.intervalSeconds = 10

最后修改 November 15, 2023: fix 404 (#71) (626ce4d)

Load balancing and health check-up

Load balancing

You can add multiple agents of the same type to the same groupto achieve the ability of load balancing.

The currently supported Agent types include:TCP,HTTP, tcpmux.

# frpc.toml
    [[proxies]]
    name = "test1"
    type = "tcp"
    localPort = 8080
    remotePort = 80
    loadBalancer.group = "web"
    loadBalancer.groupKey = "123"
    
    [[proxies]]
    name = "test2"
    type = "tcp"
    localPort = 8081
    remotePort = 80
    loadBalancer.group = "web"
    loadBalancer.groupKey = "123"

When the user connects to port 80 of the FRPS server, FRPS The received user connections will be randomly distributed to one of the surviving agents . This can ensure that even a FRPC The machine crashes, but there are still other nodes that can provide services.

For TCP type agents, it is necessary to ensure that groupKeySame for permission verification, while remotePortalso needs to be consistent.

For HTTP type agents, it is necessary to ensure thatgroupKey, customDomains(custom domain name),subdomain Same aslocations.

Health Check

By providing Agent Configuration health check parameters, it is possible to reverse the Agent When the service fails, remove the service from FRPS Remove from the middle. Combined with the function of load balancing, this can be used to achieve high availability architecture and avoid single points of service failure.

To enable the health check feature, it is necessary to add it in the Configuration of each Agent healthCheck. type={type}.

The currently supported types areTCPandHTTP

ForTCP, as long as a connection can be established, the service is considered normal.
Forhttp, an HTTP will be sent Request, the service needs to return a status code of 2xx to be considered normal.

The following isTCPExampleConfiguration:

[[proxies]]
    name = "test1"
    type = "tcp"
    localPort = 22
    remotePort = 6000
    # 启用健康检查，类型为 tcp
    healthCheck.type = "tcp"
    # 建立连接超时时间为 3 秒
    healthCheck.timeoutSeconds = 3
    # 连续 3 次检查失败，此 proxy 会被摘除
    healthCheck.maxFailed = 3
    # 每隔 10 秒进行一次健康检查
    healthCheck.intervalSeconds = 10

以下是 http ExampleConfiguration：

[[proxies]]
    name = "web"
    type = "http"
    localIP = "127.0.0.1"
    localPort = 80
    customDomains = ["test.yourdomain.com"]
    # Enable health check with type http
    healthCheck.type = "http"
    # Health check the path that sends the http request. The backend service needs to return the http status code of 2xx.
    healthCheck.path = "/status"
    healthCheck.timeoutSeconds = 3
    healthCheck.maxFailed = 3
    healthCheck.intervalSeconds = 10

Last modified November 15, 2023: fix 404 (#71) (626ce4d)