HTTP::collect performance impact

Does Collecting of HTTP request payload effect the performance of LTM?

Suppose if I collect 10K of data every request. Do I have to take any caution reagrding the performance in the case of iRule mentioned below?

Is it necessary to use HTTP::release in my case? I read somewhere that data is implicitly released.

Are there any suggestions/advice for the iRule below, so that it runs better performance wise.

when HTTP_REQUEST {
        if { [HTTP::host] contains "test-main" } {
           pool pool_test-main
           return
        } elseif { [HTTP::host] contains "test01" } {
                 pool pool_test01
                 return
        } else {
                pool pool_test02
                
                if {[HTTP::uri] contains "web_dummy"} {
                         collect
                        HTTP::collect [HTTP::header Content-Length]
                } 
                return
        }
}

when HTTP_REQUEST_DATA {
       foreach x [split [HTTP::payload] "&"] {
       if { $x starts_with "user_ref=" } {
            set user_ref [lindex [split $x "="] 1]
            if { [matchclass $user_ref equals CX_SERVICE_ID] }{
               HTTP::uri [string map {"/web_dummy" "/notify_dummy"} [HTTP::uri]]
               pool pool_test00
               return
               }
           }
       }
    HTTP::release
    return
}