In the example above you could see three different queues each one with different number of agents.

Each queue has its own configuration. Most of the settings are common, unless one option - the context.

This option is giving you the possibility to define a context, where eventually, by pressing a single digit, the caller will be "transferred". The digit could be pressed at any time while the caller is waiting in the queue. The context we are talking about has to be created in the extensions.conf configuration file. There is one more requirement and it is that you have to create an extension of one digit, in this context - the digit that the caller will press in order to exit the queue.

Despite of the fact, that for all the queues, the setup for leaving a number and then emailing it to our email address will be pretty much the same, we recommend you to create different contexts for every different queue. That is because we are going to email not only the left number but also some other information such as the queue from which the caller has quitted. Of course you could do it with one common and more complicated context for all the queues, but we think that it will be much easier to manage and change the settings for each queue independently of the other ones.

So here are all the options in our queue and what they are doing"

First of all we have three extensions for the incoming calls. They are all put in the [incoming_calls] context. Their purpose is to put the incoming calls in the corresponding queue.

When we have an incoming call, the first thing we are doing is to change the Callerid name to the name of the queue where the caller will be put. Thus the agent who is answering the call will know from which queue exactly the call is coming. It is not absolutely necessary, but the idea is that one agent could be assigned to many different queues. In order to do the change we are using the Set application - More information about it in our tutorial

The next step is to actually send the call in the queue - pretty easy one. There is an application called Queue to which you just have to pass the name of the desired queue as an argument. We are not going to use any extra options. You could find more information about how to use the QUEUE application in our tutorial

You could notice and the usage of the Hangup application. It is always a good idea to use it as last application for all of your extensions. Thus, you will always be sure that the used channel will be released when the conversation is over. Refer to our tutorial for more information on this application

Next step is to create the contexts, where the callers should be "transferred" in case they have decided to exit the queue. As we have mentioned above we have different context for each of the queue.

Before we continue - a few words about the idea behind. We are running queues with a callback option. In other words, when we have a caller waiting in the queue, we will play him a message every 60 seconds. The message will saying: "All of our lines are busy. You could leave your number by pressing one and we will call you back as soon as possible or you could stay in the queue and wait to be served.".

If the caller decides to exit the queue, he will be asked to leave his number after the "beep" signal. At this stage, you have to keep in mind that the caller might not leave his number for some reason. So what we will have a check and if the number is not left we will send an email message that the caller has decided not to leave his number.

So what do we need as extensions? The first thing we are doing is to set a flag. Why? In case the caller hangs up at some point before leaving his number, we need to send a message that we did not received the number. For the purpose we are using the so called predefined extension - h, which allows you to execute something in case of hang up. Refer to our tutorial for more information

However, there is a slight chance that the caller might wait for the timeout to expires without leaving his number or to have a problem with the sending of DTMF tone, that we will have an email saying that the caller has left his number but the number field will be empty and because we have a hang up event even if the caller has left his number, we need the flag to determine whether the hang up is before the application that will store the number or after it and on that basis we will know whether we need to send an email message on hang up or not. In this way we omit the sending of one message twice.

On the next step we have the Playback application. Its simple purpose is to play a sound files. We have a sound message, which instructs the caller to type his number after the beep tone and to press the pound key (#) when he is ready. For more information refer to our tutorial about the Playback application

Next, comes the Read application. It’s the one used to store the number typed by the caller. It has many options.

There is a small limitation in the application we are going to use to save the number typed by the caller. We need to know, when he has finished with the typing. So there are two ways.

The first one is to ask him to press the pound key (#) once he is ready and the second one is to limit the length of the number the caller could type plus a timeout.
In our example we are going to limit the length of the number to 10 digits and we are going to put a timeout of 5 seconds. We are going to store the number typed by the caller in the NUMBER variable. The Read application gives you the possibility to play a sound file before the typing of the number. In our case we will are going to play a beep tone

Refer to our tutorial for more information and details about the Read application

Next steps - we have the Wait application, which is not absolutely necessary and then we set the flag variable to 2. This, as we have already described above will be used to omit the sending of one and the same email message twice. For more information about the Wait application, refer to our tutorial.

We now have to check whether the caller has actually left his number or not. For the purpose we can use the GoToIf application. It has a very specific syntax, but its idea is simple to verify whether one condition is true or false. So, in our case we are just checking whether the NUMBER variable is empty or not. If it is empty then the execution of the dialplan will continue with the next step, marked as n(empty). n makes your life easier as you do not have to write and follow the steps (priorities) - n means that you take the previous priority and increase it with 1. (empty) is a label showing to you and the system to which n you will be send.

If the NUMBER variable contains even one digit then we assume that we have the number of the caller so the execution of the dialplan will be send to the step with label (full). Refer to our Tutorial for more information, details and option about the GoToIf application.

Both steps (empty) and (full) have one and the same task - to execute the shell script, which will form the email message and send it. The only difference would be in the body of the email message. The application, which we have to use is called System. It just executes system commands as you will do it in your Linux CLI. Refer to our Tutorial for more information about the System application.

In our case we are executing our own scripts called mailnumber and mailnonumber. They are written in such way that we have to pass them a few parameters. You could see them above as screenshots and in pure text

If we do not have the caller's number, we will execute the script called mailnonumber. We need to add, as first parameter, the email address of the recipient. The second parameter is the CallerId number as we have received it in our system. The last parameter is the name of the queue from which the caller has exited.

If we have the caller's number we will have executed another shell script called mailnumber which has a different body, but the same parameters plus one more, which of course is the number left by the caller.

The sending of the email is the last step we need to do and that is why after it we have to hang up the channel. For the purpose we will use once again the Hangup application.

In the predefined extensions starting with h we have a simple check of the FLAG variable to help use determine whether we have to send an email with saying that we do not have the caller's number. (this step is only needed in case the caller hangs up before the prompt to leave his number)

They are not very complicated. All you need is a basic knowledge of shell scripts and the mime-construct package. apt-get install mime-construct (in Debian) and you will have it. Then its just up to the mime-construct's syntax