Route to different pools based on /content

Chris,

The string first command works nicely with this because you can skip the first forward slash and get the index of the second forward slash (after the /qa or /uat directory) using the "startIndex" parameter. But the string range commands you used have the same effect. But make sure you're checking for URIs starting with a forward slash in your switch statement (/uat and /qa). And you'll need a -glob flag on the switch statement.

Greg,

The HTML content that the server replies with won't have these /qa or /uat prefixes unless the servers have been configured to include them. You could try to insert them in all of the HTML, but it would get complicated trying to parse the HTML. If the initial request from the users contains the prefix, you could set a cookie to track that they're a UAT, QA or general user. You could check for that cookie if the URI doesn't match a UAT or QA URI.

Here's an untested example:

when HTTP_REQUEST {
    Default to not setting a pool cookie
   set pool_cookie ""
    Check URI set to lower case with wildcard matching
   switch -glob [string tolower [HTTP::uri]] {
      "/uat*" {
         pool uat_pool
         HTTP::uri [string range [HTTP::uri] [string first "/" [HTTP::uri] 1] end]
         set pool_cookie uat
      }
      "/qa*" {
         pool qa_pool
         HTTP::uri [string range [HTTP::uri] [string first "/" [HTTP::uri] 1] end]
         set pool_cookie qa
      }
      default_pool {
          Check if user has pool cookie
         switch [HTTP::cookie pool_selector] {
            "qa" {
               pool qa_pool
            }
            "uat" {
               pool uat_pool
            }
            default {
               pool general_pool
            }
         }
      }
   }
}
when HTTP_RESPONSE {
    Set a pool selector cookie if a UAT or QA pool as selected in request
   if {$pool_cookie ne ""}{
      HTTP::cookie insert name pool_selector value $pool_cookie
   }
}

Aaron