Hi,

I'm Joe. I'm a product manager in JSD cloud. I'll be mainly working on the knowledge base. Would any of you be interested in getting on 20-30 min call with me? I'd like to understand this need more and also understand how your organization/team manages knowledge base. 

You can email me at jjayanth@atlassian.com and I can schedule some time with you. 

cheers

joe

And it seems that non licensed confluence users can exist in a group in confluence. As long as they are not part of a group with global permissions to use the site and Unlicensed Access is granted from SD to KB space.

Cool Guys! Now the Organization and Group alignment is the only real big hassle I guess. 

A workaround, at least to start with, could be to allow Service Desk customers in Confluence without them costing a license? The use case is the following.

Use Case

One Service Desk project serving multiple customer organizations. The customer organizations each have KB articles the other ones can't see and some are common. My customers then ask me if you can connect multiple KB spaces because of this requirement. 

Workaround

A solution I have tried, on server, using one KB space but that seems support the above scenario is the following. 

• One common service-desk-customers group. All service desk customer belong to this group. This group is given the role of service-desk-customer on the SD project.
• A customer specific group for each customer, service-desk-customer-a, service-desk-customer-b, etc. 
• The KB space has a top level, which contains KB articles common to all customers. There are no page restrictions on these articles.
• Each customer organization has a page under the top level. All customer specific articles are placed under this page. 
• A view restriction is placed on this customer specific level page corresponding to the customer specific group. Now only customers belonging to this group can view the pages.
• Service-desk-agents group is given edit rights on all pages.

Now Agents can create KB articles and place them where needed. Common will be above all customer specific pages in the tree. Obviously, customer specific articles go under their respective branch in the tree. 

Logging into Portal as a customer who is in group service-desk-customer-a and only KB articles which are common and under the customer page where they have view right based on the group show up in portal.

This seems to work fine. One SD project with one KB where articles are view-able based on the structure and page view restrictions.

The Problems

• Service Desk customer now cost a Confluence license
• The Organization custom field in JIRA SD needs to be kept in sync with the real directory groups. Would be nice if we could actually use real groups for Organization in SD. 

 Suggestions

If all customers automatically go into a service-desk-customers group and all organizations automatically create a group and place customers in the corresponding group and keep it updated.(nested groups) Then treat the service-desk-customer group special in Confluence so they do not cost a license then the above workaround is pretty much supported.

Also this make me think whether multiple KB spaces and one SD project is the right way to go. This workaround points more to segmenting a single KB space. For example.

I create a new organization for a new customer. You could then give the option of adding a "area" in the KB space. If the agent says yes a top level KB page is created in the KB space and the view restrictions are applied. When creating new articles the agent simply selects which org to place it under or whether it is common. Now we don't have many KB spaces cluttering up Confluence.

Hi,

Thanks for your feedback!

We have multiple specific issues to help us better track the requirements. Please vote for the suggestions that are important to you:

• JSD-170: link multiple Confluence spaces to a single Service Desk project
• JSD-35: link Confluence Questions as a knowledge base for a Service Desk project
• JSD-3480: display known issues (customer-visible JIRA issues) on the customer portal
• JSD-143 (this issue): link knowledge sources that are not Confluence spaces / Confluence Questions. Please also add a comment mentioning specifying which knowledge source you want to integrate with.

Breaking up bigger feature suggestions into smaller, more specific ones helps us better understand the requirements and prioritise the features, as they often have very different solutions and implementations.

Thanks!

-Shihab

It'd be nice for us if we could search multiple Confluence spaces for certain request types, but I'm particularly interested in being able to show "known problems" on the main customer portal and/or on the individual project portals, similar to what Atlassian does in the own instance.

I might prefer to show the list at the top/bottom of the page rather than in a popup, but either way would be a huge benefit. The problems would fall into three main categories:

• We know there's a problem, we're working on a solution
• We know there's a problem, here's a workaround, we're working on a proper/permanent fix
• There was a problem; it's fixed now, but you might need to do something (such as restart your application, power-cycle phone, etc)

I'd envision this working by specifying a filter for each portal, such as

issuetype in (Bug, Problem, Outage) and (statusCategory != Done or resolutionDate > -1d) order by Severity DESC

. Then the ability to create "Quick Filters" to place issues into groups:

And then a place to specify a max number of issues to display.

If users could "Watch" the issues right from the list, that'd be nice to, but maybe that's for phase 2. Or to make it more generalized, they could be added to a new group of "Affected Users" which admins could reference in their notification schemes and workflows to have full control over when notifications are sent.

@shihab Hamid, we also need this feature. I have searched and found at least 3 related issues for this request. I'm sure there are more. Can you please consolidate them into one to get an accurate number of votes and advise when this will be available? Thanks

1. What types of knowledge bases do you have in your organisation?

• Multiple Confluence Spaces
• Confluence Questions

2. One of the challenges with searching multiple sources is combining results for heterogenous systems - as each system independently produces a result for the "most relevant articles" in that system. How would you like to combine results? Is there one system that should take priority? (what system is that?) Or would you like to see the first 3 results from each system?

• Not applicable as we're only using Confluence - happy with the current way Confluence handles search. Would require a note saying which space/area the article is located to help guide

3. Do these knowledge sources require authentication / end users to log in before they can search for knowledge articles? If so, how does the authentication work?

• Logins to JIRA & Confluence. Access would be controlled based on these logins (i.e. group membership)

1. What types of knowledge bases do you have in your organisation?
Multiple Confluence Spaces

2. One of the challenges with searching multiple sources is combining results for heterogenous systems - as each system independently produces a result for the "most relevant articles" in that system. How would you like to combine results? Is there one system that should take priority? (what system is that?) Or would you like to see the first 3 results from each system?

Combine the output of the search results. List in order of most relevant results.

3. Do these knowledge sources require authentication / end users to log in before they can search for knowledge articles? If so, how does the authentication work?

All our customers have logins to JIRA / Confluence. Their logins would be configured to have permission to each of the knowledge bases.

1. What types of knowledge bases do you have in your organisation?
Multiple Confluence Spaces

2. One of the challenges with searching multiple sources is combining results for heterogenous systems - as each system independently produces a result for the "most relevant articles" in that system. How would you like to combine results? Is there one system that should take priority? (what system is that?) Or would you like to see the first 3 results from each system?
Either approach would work for us

3. Do these knowledge sources require authentication / end users to log in before they can search for knowledge articles? If so, how does the authentication work?

No authentication is required

Here is from us:
1. One Confluence with multiple spaces, need to be able to look on the entire instance, limit it to one or more spaces. If this would support the space hierarchy plugin it would be even better. Not being able to look inside the entire instance makes the integration almost useless. Yes we do have Questions too, still evaluating it, hard to find any reason to buy price / question would be astronomical.

We do have a GSA and we do work on trying to implement a JavaScript injection into confluence search, so it will bring answers from outside confluence. If you do support GSA directly it would be great.

2. Scores are always numbers between 0..1, allow user to configure a multiplicator for each search source so you can combine the results correctly. An admin can tune the system to find the proper values of the search source multiplicators (again numbers between 0..1) in order to obtain the desired results.

3. We do have authentication for our sources but most content is freely accessible to anonymous users. It would be wise to postpone authenticated search for a further release. For us, we do think that we would get 95% of the usability without authentication support. (Basic HTTP auth should do the trick)