Allowing external access to Processor

[David King]

Hi all,

I'm looking at setting up a processor that allows shorter URL's to KB articles. I was wondering if there's a good way to allow external access (Via ACL?) to a specific processor?

Background

Links to our KB articles currently look like:

https://instancename.service-now.com/csm?id=kb_article&sys_id=7abc123odadoiwagbdwaiohvsuicpoe6

I've got a processor which shortens the links to:

https://instancename.service-now.com/help?article=KB00001

The marketing department will be much happier pasting this into their social media links!

The Processor

It's a relatively simple thing that executes the following:

(function process(g_request, g_response, g_processor) {
	
	var article = g_request.getParameter("article");
	
	//Query table for Knowledge article
	var kb = new GlideRecord('kb_knowledge');
	kb.addQuery('number',  article); //Key, query value
	kb.query();
		
	//Default url if not found
	var url = 'https://devinstance.service-now.com/csm';
	
	//While we have results
	while (kb.next()) {
		if (kb.number.toString() == article.toString()){
			url = url + '?id=kb_article&sys_id=' + kb.sys_id.toString();
		}
	}
	//Redrect to base csm addr or found kb article
	g_processor.redirect(url);
		
})(g_request, g_response, g_processor);

The CSM endpoint is accessible by external users, they don't need to be logged in to navigate the service portal and read KB and News articles.

However, this processor is set up in the global namespace with the previously mentioned /help endpoint.
Navigating to <instance.com>/help requires a user to log in.

Is there a safe way to enable access to this specific endpoint for external users?
Any comments are appreciated!

Regards,

David

[Goran WitchDoc]

For fun, if you just try to redirect to /csm and not anything with knowledge etc. does it still redirect to SSO. Just to verify it isn't something else with the portal.

[David King]

Good call, I just realised that I don't believe I've tried accessing /help.do as my admin account. I previously just attempted it with the public user (Not snc_external as I now know! Ironically it's an internal role for accounts to mark that they're external from the business, as opposed to external from the system, which is public)

I noticed that I'd forgotten to add the public page, so I've activated that module and added an active record in Global for "help" (I dont believe the public page entry requires a .do)

Now a public user can access the endpoint without having to log in!

But the result is a "page not found" for help.do

[David King]

Ah, forgive me again, it seems that all references to the "help.do" endpoint need to be named "help" and then just referenced as .do when accessing it outside, it's now working and avoiding the SSO!

Thank you so much for your help, I'll be sure to poke around for hidden records when trying to tweak the instance next time, as mentioned before, if you do make the video please share it with me, would be great to see!

Thanks again,

David

[Goran WitchDoc]

Btw. here is the video 😃

[David King]

Nice video! I'll keep an eye out for more of them on your channel 🙂

As a small extension of this processor (The marketing team demanded more than just the KB Article number, ho hum) I've added a custom field to the KB Article table called "Friendly name", which allows the marketing team to 'add their own urls'.

Plus a small sanity check for internal / external articles. If a non logged in user tries to view an internal article then the ACL's will restrict their access, however it's a little bit neater to redirect them to the /csm page instead of a "Article not found" page.

The code now looks a little like this:

//g_response.setContentType('text/html;charset=UTF-8'); // Not necessary here, but may serve use in future

(function process(g_request, g_response, g_processor) {
	//
	// Be aware, debugging with writeOutput will nullify the effects of a redirect
	// Remember to comment them out when done debugging
	//
	
	//Hard coded example
	//g_processor.redirect('https://instancename.service-now.com/csm?id=kb_article&sys_id=7070431f4f9f5b009417d6ef0310c726');
	
    
    // Parameters
    var article = g_request.getParameter("article");
    var friendly_name = g_request.getParameter("help");
    
    //Find external knowledge base id
    var kb = new GlideRecord('kb_knowledge_base');
    kb.addQuery('title', 'External Knowledge Base');
    kb.query();
    var knowledge_base_id;
    while (kb.next()){
	if (kb.title.toString() == 'External Knowledge Base'){
	    knowledge_base_id = kb.sys_id;
	}
    }
	
    //Catches 404's -- Shouldn't occur, but good practice
    var url = 'https://instancename.service-now.com/csm';
    
    if (!knowledge_base_id){
        g_processor.redirect(url);
    }

    var kba;
    if (article) {
        
        //Query table for Knowledge article
        kba = new GlideRecord('kb_knowledge');
        kba.addQuery('number', article); //Key, query value
        kba.query();

        // Make sure attributes match
        while (kba.next()) {

            if (kba.number.toString() == article.toString()){
                if(kba.kb_knowledge_base.toString() == knowledge_base_id.toString()) {
                    url = url + '?id=kb_article&sys_id=' + kba.sys_id.toString();
		    g_processor.redirect(url);
                }
            }
        }
    }
    

    if (friendly_name) {
        // For friendly urls
        //Query table for Knowledge article
        kba = new GlideRecord('kb_knowledge');
        kba.addQuery('u_friendly_url', friendly_name); //Key, query value
        kba.query();

        // Make sure attributes match
        while (kba.next()) {
            if (kba.u_friendly_url.toString() == friendly_name.toString()){
                if(kba.kb_knowledge_base.toString() == knowledge_base_id.toString()) {
                    url = url + '?id=kb_article&sys_id=' + kba.sys_id.toString();
		    g_processor.redirect(url);
                }
            }
        }         
    }		

    //Redrect to base csm addr or found kb article
    g_processor.redirect(url);
		
})(g_request, g_response, g_processor);

There's alot of duplicate code there, but  it was a quick and easy copy and paste job that ended up working, I had tried playing with extracting the GlideRecord queries to a separate function called findRow() but the results when passing back a GlideRecord object were a little interesting, it returned null rows even when calling GlideRecord.next() in the function

Maybe I'll tidy it up on day, but this seems to be a good functional start! (Not fully tested, that's today's job 😉 )

Thanks again for your help!