Question

I have an IIS reverse-proxy delivering an application with a RoR 3 backend. The application displays information from our databases over a period of time chosen, giving you various charts and graphs. When not run through the reverse-proxy, everything runs properly for whatever time period is shown. But when run through the reverse-proxy, the response body will just get cut off if enough information is being sent.

Looking at the rails backend, all of the information is being sent out. On the browser side I can look at the response body to see where it suddenly cuts off. If I use a small window of time, the full response will be sent and work fine, but once it gets to a certain size range, it stops working properly. The cutoff point will also change, even if I’m looking at the same time range.

So, why is the response body not being fully sent by IIS? Or what controls that?

Answer 1

I was just having this problem with an Apache/Tomcat backend server. I found this solution, which worked perfectly for my application:

You need to increase ARR Response Buffer Threshold. Please follow
these steps:

1. Open IIS Manager, navigate the server level.
2. In the Features View, double-click the “Application Request Routing Cache” feature.
3. Click Server Proxy Settings… on the Actions pane.
4. On the Buffer Setting section, set Response buffer threshold to a proper value according to your needs.

(Source)