What is a reverse proxy?
A reverse proxy is an application that receives requests from the Internet and forwards them to a set of servers. These servers are usually located on an internal network and are not directly accessible from outside.
Reverse proxy capabilities:
- Load Balancing (TCP Multiplexing)
- SSL Offload/Acceleration (SSL Multiplexing)
- Content Switching/Redirection
- Application Firewall
- Server Obfuscation
- Single Sign-On
Why use a reverse proxy with OutSystems?
- Reduce Load on Application Servers
Reverse proxies have the ability to compress and cache content, encrypt data (HTTPS), relieving these tasks from the application server.
- Increased security and traceability
With all application traffic passing through the reverse proxy, logging, authentication and access control can be configured and managed in a centralized place. It also prevents Internet exposure of application servers, protecting from vulnerabilities the servers might have in other software or services.
- Ensure High Availability
Reverse proxies can support high-availability methods. This will allow you to eliminate downtime. In an OutSystems farm scenario, there are multiple application servers. The reverse proxy will then enforce a load balancing technique like round-robin to distribute the load among the servers in the cluster.
When a server goes down for maintenance, the system will automatically failover to the next server in the round-robin sequence keeping the applications available.
When serving OutSystems applications through a reverse proxy, the following requirements must be guaranteed:
- Maintain original host header of the request
The original host header must be sent to the application server. OutSystems uses this name when generating URLs that go in the HTML. Keeping the original host header prevents internal server names to be used in the HTML generation and ensures URLs are correct.
- Referenced content availability
All modules that are used somehow by your application must be available from the outside world. For example, if you have a module myApp that uses a screen in myOtherApp module for logging in, and images from an ImageRepository module, then you need to have the following URL available to the outside world:
- Only the default ports for protocols HTTP and HTTPS are supported
OutSystems only serves applications in HTTP port 80 and HTTPS port 443.
Additionally, when the back-end of one or more OutSystems mobile apps is hosted behind a reverse proxy, the following requirements must also be met:
- Do not cache mobile app resources in the reverse proxy
The loading of mobile app resources is already optimized. In most cases, these resources are loaded only once. Make sure that the reverse proxy does not cache them to prevent runtime problems in your mobile apps.
- When using keep-alive/persistent connections, set the connection timeout value to at least 30 seconds
If your reverse proxy is using persistent connections and thus issuing keep-alive headers, make sure the connection timeout is set to at least 30 seconds. If you are using Apache (or an Apache-based) web server being used in reverse proxy mode, review the keep-alive configuration settings being used, since in these servers the default connection timeout value is 5 seconds. This configuration adjustment is described in F - Adjust keep-alive connection timeout.
- Applications available externally using the hostname configured in Service Center and via HTTPS
Applications must be externally accessible using the hostname configured in Service Center and served via HTTPS.
- URL paths of applications must be kept unchanged
The reverse proxy must not transform URL paths of applications in any way, i.e. applications must be externally available at https://<hostname_in_Service_Center>/<application_name>