Queuing client requests on bigip

Question

Hi 
&nbsp;  
&nbsp; I need to solve the following issue.  
&nbsp; Say I have 1000 simultaneous requests (POSTs) to my single webserver which can process - say - 10 requests per second. Is there a way to put the 990 remaining requests in a queue to spoon feed the server? 
&nbsp; What is the easiest way to accomplish this sort of buffering? 
&nbsp;  
&nbsp; Thanks 
&nbsp; Filip 
&nbsp;

hoolio · Answer

Hi Filip, 
&nbsp;  
&nbsp; I don't think there is a safe method to queue HTTP requests on LTM.  This has been discussed a few times.  This post has some possible options: 
&nbsp;  
&nbsp; http://devcentral.f5.com/Default.aspx?tabid=53&amp;forumid=5&amp;tpage=1&amp;view=topic&amp;postid=25174 
&nbsp;  
&nbsp; Aaron

colin_walker_12 · Answer

The nearest to this that I've seen done wouldn't work with POSTs, only GETs.   
&nbsp;  
&nbsp; The method was to count the connections to the server and when the threshold was reached, insert a custom cookie then  use HTTP::respond to send back a "please hold" type queuing page to the user that had a built in meta refresh.  The cookie could then be used to prioritize those that have already been queued (or those that have been queued more than x times, or whatever you wanted), allowing them in first when the server had more free connections. 
&nbsp;  
&nbsp; Colin

Forum Discussion

SSO Login Update Coming to DevCentral

Kevin - June 2026 Featured F5er

Tyler - May 2026 Featured Member

Recent Discussions

Ivanti MDM Core & F5 LTM/ASM with mTLS

F5 BIGIP & XC certbot plugin

F5 radius login not working

A Method for Auth and SSO

setup F5vpn using key stored in TPM?

Related Content

Using Client Subnet in DNS Requests

HTTP Request Queuing for WAM

Implementing Client Subnet DNS Requests

Implementing Client Subnet in DNS Requests

HTTP Request Smuggling Using Chunk Extensions (CVE-2025-55315)

ABOUT DEVCENTRAL

RESOURCES

SUPPORT

PARTNERS