Configuration

Configuration files

Internal configuration

JLupin Edge Balancer internal configuration consists of the following files:

• NGINX configuration file nginx.conf located in $JLUPIN_HOME/platform/technical/nginx/<os_type>/conf which only includes edge.conf - the main JLupin Edge Balancer user's configuration file.
• Virtual servers definition: data, admin and discovery located in $JLUPIN_HOME/platform/technical/nginx/<os_type>/conf/servers
• Services definitions that are individually included by virtual server and are located in $JLUPIN_HOME/platform/technical/nginx/<os_type>/conf/services.

The list of available service is as follows:

• common - basic parameters that are common for all types of virtual servers.
• proxy - dynamic proxy for HTTP request to microserivces with healthchecking included.
• discovery - discovery services that is used for intra-zone routing
• ssl - SSL support, it can be applied to any virtual server except discovery one (that provides /_discovery URI).
• status - information about node, microserivces and edge balancer itself.
• debug - specific set of entry points that shows the internal state of balancer, if you know the internal architecture of JLupin Edge Balancer it could be useful in troubleshooting (before going in to debug mode).
• addons - additional services that could be useful in development and/or maintenance (for example: swaggerUI).

Each virtual server type includes specific list of services:

data

• common
• proxy

admin

• common
• proxy
• ssl
• debug
• discovery
• status
• addons

discovery

• common
• discovery

NOTE: These files SHOULD NOT be edited by the user unless JLupin support allows it (for example, while providing appropriate fix for the bug). If some of the parameters provided by above files should be changed, it should be overwritten in user's configuration.

User's configuration

The main JLupin Edge Balancer user's configuration file is called edge.conf and is located in $JLUPIN_HOME/platform/start/configuration. It contains:

• Basic configuration of NGINX server (generic parameters)
• JLupin LUA module configuration and initialization commands
• 'admin' and 'discovery' virtual servers.

It also refers to additional configuration files / directories:

• $JLUPIN_HOME/platform/start/configuration/edge_servers directory, where definitions of 'data' virtual servers are located. This is the place, where you can add / modify configuration of access points for your services provided by microservices.
• $JLUPIN_HOME/platform/start/configuration/edge_custom directory, where additional configuration files are located, including external microservices.

Parameters

Common parameters

Virtual servers parameters

Each virtual server that provides services from servlet microserivces is defined through server directive and configured by the following parameters

The additional generic NGINX parameters can be provided by the user to adjust Edge Balancer actions to services' specifics (offical documentation).

External microservices parameters

External microservices' configuration is a part of custom configuration and is defined in $customConfigDir/external_microservices.lua file. It contains the list of microservices that are managed outside JLupin Platform, but they should be provided and load balanced by JLupin Edge Balancer. There is only one requiremet towards external microservice - its services should be provided by HTTP protocol on the port available to JLupin Edge Balancer (from network perspective).

The list is a LUA table definition, each entry is a external microservice definition, as show on the below example:

externalMicroservices = {
  {
    isActive = false,
    name = "tomcat9",
    contextName = "/docs",
    directPort = 8080,
    externalPort = 8000,
    externalHealthcheckURI = "/isAlive",
    isExternalHealthcheck = false,
    httpStickySession = false,
    httpStickySessionCookieOptions = ""
    -- host = "127.0.0.1" -- If not set it is equal to discoveryHost defined in edge.conf
  }
}

Parameters description:

Failover

By default Edge Balancer performs full failover for HTTP 504 and 502 errors, which means that there is possibility that some requests which were processed during the failure would be processed again. To prevent such case the proper mechanism protecting system against request duplication should be provided on the application level OR the following configuration should be provided:

on Windows:

balancerHttp504Failover = 1
balancerHttp504FailoverRetries = 3
balancerHttp504FailoverCheckProcess = 0
balancerHttp502Failover = 0
balancerHttp502FailoverRetries = 3
balancerHttp502FailoverCheckProcess = 1

on Linux:

balancerHttp504Failover = 1
balancerHttp504FailoverRetries = 3
balancerHttp504FailoverCheckProcess = 0
balancerHttp502Failover = 0
balancerHttp502FailoverRetries = 3
balancerHttp502FailoverCheckProcess = 1

and additionally the following rules should be applied to the iptables configuration:

iptables -A OUTPUT -p tcp --sport 30001:40000 --tcp-flags RST RST -j DROP

where 30001:40000 reflects the scope of available ports for SERVLET microservices defined in main.yml file. It should be changes accordingly to your configuration (if it has been changed).