Revision

Date	Author	Version	Description

Date	Author	Version	Description
Jan 18, 2023	@Sharmendra Vishwakarma / @Gaurav Gurudas Nadkarni (Unlicensed)	1.0	Analysis and solution description along with WBS, Risk and assumptions

1. Problem Statement

Customer is using BRIKIT app on confluence server. Recently, they decided to migrate from confluence server to confluence cloud after which they started experiencing layout related issues in the cloud. The pages which were using BRIKIT plugin and its associated themes and components were having issues. In most of the cases the components were no visible at all on the screen.

2. Analysis

The main issue here is that the Brikit plugin is only available for server/DC but not cloud. Hence, when we are try to migrate the pages, the pages are not showing in Confluence.

2.1 To Do

Need to analyze and check what are the possible option along with the solutions to migrate the BRIKIT develop confluence page to confluence cloud

2.2 Pre-requisites

Sl No#	Items	URL	Description	Status

Sl No#

Items

URL

Description

Status

1

Access to confluence cloud

https://rakutencartera-sandbox-771.atlassian.net/wiki/spaces/CC/pages/2694302/Customer+Care+Org+Chart

https://rakutencartera-sandbox-771.atlassian.net/wiki/spaces/CC/overview?homepageId=2687141

For Gaurav

open

2

Access to confluence Server

For Gaurav

open

3

cloud app that can render HTML

https://marketplace.atlassian.com/apps/252/html-for-confluence?hosting=cloud&tab=overview

2.3 Analysis

Things we have tried so far

Things we have tried	Observation(As things stand)	Status

Things we have tried	Observation(As things stand)	Status
Exported the pages(with BRIKIT macros) from confluence server and imported them to cloud in the word format	It did not come up good. Alignment were not proper. Many of the components were missing.	Not satisfactory
Exported the pages(with BRIKIT macros) from confluence server and imported them to cloud in the pdf format	It did not work	NOT SATISFACTORY
Trying with apps which are already existing in marketplace to do this	We tried the following apps -: https://marketplace.atlassian.com/apps/252/html-for-confluence?hosting=cloud&tab=overview We do not think it can be achieved using marketplace app. These are mostly HTML, XHTML and XSLT macros. BRIKIT uses some custom macros and we do not think we can migrate the BRIKIT custom macros using these apps.	NOT SATISFACTORY
Reaching out to Atlassian	Atlassian were not much keen as this was an app related issue	Not much of help
Reaching out to Vendor(BRIKIT) for help	Yet to reach out to the Vendor but we do not think it will help us as they themselves do not have any app for cloud and may not be interested in sharing their knowledge with us about inner workings of their server app	Not much help expected
Think about doing this using scripting(Server to cloud)	Scripting will also have issues. Some issues are as follows, Confluence cloud api has very little help as far as creation of content is concerned. There are many parameters which we can pass to api but there is no much clarity on what these parameters do. Automating this process itself will take time as we will have to write the logic to place the components on the screen using apis	DIFFICULT to Archieve
Think about doing this using scripting(Server to Server) using REST API	Using scripts to convert pages which uses BRIKIT layout to pages which do not have BRIKIT layout in them and then using normal server to cloud migration to migrate these new pages which do not use BRIKIT components Using REST API	Worked
Manual creation of pages	This should be our last resort	Out of contention
Using Automation(Selenium)	We are in the process of trying out this approach	didn't work

2.3.1 Approach Using Automation(Selenium)

We are planning to use automation scripts using selenium to migrate the pages and spaces from Rakuten confluence server to cloud. The idea behind this is to use the capability of the selenium script to automate the interaction which user can have with confluence server. Using this approach we can download the data from the confluence server pages(scraping data from the screen or any other method) and store it locally along with the attachments in those pages. The ideal steps for implementation could be as follows,

Selenium will be used to script the automation to go to different pages in different confluence server spaces
Selenium script will then scrap the data from the screen
Selenium script will also download the attachments from the different pages
Selenium script will then store the data scraped from the screen in some data format ex. excel or csv.
We are also planning to use Selenium to create migrated spaces and pages in the confluence server itself. These migrated spaces and pages will have the same data from the actual pages on confluence server but without any BRIKIT components or macros
Team will decide on a generic layout which we can use for these migrated pages
We will push the data that we scraped/downloaded using selenium script from confluence server to migrated spaces/pages on the confluence server itself using selenium scripts
This way we will have spaces/pages in the confluence server which do not user BRIKIT components or macros but contain all the data from the original spaces and pages
We can then do the migration of the confluence(server to cloud) again and then we will have two sets of spaces, one with BRIKIT components and macros and other which are newly migrated without any BRIKIT components or macros
Once the migration to cloud is done, we can delete one set which has all the spaces(and pages under them) which use BRIKIT macros. After this we will be left with spaces and pages which do not use BRIKIT components or macros
For the links which are embedded in the confluence server pages, we may have to reach out to Atlassian to point them to correct resources on the confluence server
We will be using Selenium scripts to upload the attachments to different pages as well

2.3.2 Approach Using Server REST api to convert BRIKIT pages to non-BRIKIT pages on the server itself

Using the REST api create the space and append Migrated in front of the title of the space
Using the REST api create the page under the migrated space where we will move the contents of the original page
Using the REST api pull in the details of the page to be migrated
Read the body from the REST api response and filter out the BRIKIT related classes and replace these with new non-BRIKIT(simple HTML) elements
Push this new content under the newly created page under the new(migrated) space and then move on to the next page or space

#	BRIKIT Macro	HTML/CSS Replacement	Remarks

#	BRIKIT Macro	HTML/CSS Replacement	Remarks
1	list-children
2	content-column
3	content-block
4	content-layer
5	action-button
6	html-wrap
7	content-layout
8	rotate-content
9	lesson-plan
10	catalog-navigator
11	block-link
12	html-link
13	list-attachments
14	rollover-image-link
15	restrict-content
16	target-new
17	page-turner
18	html-clear
19	display-page-title
20	display-space-name
21	display-space-description
22	choose-timezone
23	hello
24	target-current
25	include-content
26	list-spaces
27	list-favorites
28	html-copy-element
29	html-add-class
30	comment
31	blog-tree

API Details(Confluence Cloud)

	Requirement	API	Body Data	Description

Requirement

API

Body Data

Description

1

Adding content to the page(simple text content)

https://host/wiki/rest/api/content/{contentid}

{
    "version": {
        "number":1 
    },
    "title": "Testing",
    "type": "page",
    "body": {
        "storage": {
            "value": "<h1>Problem Statement</h1><p>Customer is using BRIKIT app on confluence server. Recently, they decided to migrate from confluence server to confluence cloud after which they started experiencing layout related issues in the cloud. The pages which were using BRIKIT plugin and its associated themes and components were having issues. In most of the cases the components were no visible at all on the screen.</p><h1>Analysis</h1><p>The main issue here is that the Brikit plugin is only available for server/DC but not cloud. Hence, when we are try to migrate the pages, the pages are not showing in Confluence.</p>",
            "representation": "storage"
        }
    }
}

Version: The version of the page

title: title of the page

page will be created like this

2

Adding attachment to the content

http://myhost/rest/api/content/{contentId}/child/attachment

curl -D- \ -u admin:admin \

-X POST \ -H 'X-Atlassian-Token: nocheck' \

-F 'file=@"example.txt"' \

-F 'minorEdit="true"' \

-F 'comment="Example attachment comment"; type=text/plain; charset=utf-8' \

Added files will be able to seen in this place.

API Details (Confluence Server)

#	Requirement	API Detail

#	Requirement	API Detail
1	Get spaces and pages under those spaces
2	Get content of the page
3	Update content of the page

2.3.3 Conclusion

We tried both Selenium and API approach to extract the confluence data. Using selenium we are unable to scrape the data fully along with the attachments. Eventually decided to use REST API to extract , convert and save the data in confluence server. Once conversion is done, admin needs to migrate the converted pages to confluence cloud

3.Solution Description

3.1 Pre-requisites

Sl No#	Items	Description

Sl No#	Items	Description
1	Know how many spaces are available in confluence server
2	Know how many pages are available in confluence server
3	Create new instance and copy all the spaces and pages in Confluence server
4	Need access to confluence Server to run the conversion script	Permission to install the software( python )
5	Create Brikit element conversion mapping sheet	Prepare list of mapping between Brikit element and HTML element

3.2 Solution

Flow Diagram

Approach

Steps	Step Description	Remarks

Steps	Step Description	Remarks
1	Copy all the spaces and pages from the confluence server [ Without Script]	This step should be done without any script
2	Data Extraction [ using Script]
2.1	Get the list of all Spaces and pages recursively
2.2	Go to each space, read the page and extract the content
2.3	Check if the page is using brikit theme
2.3.1.1	If yes , then relace the brikit elements with HTM by applying the conversion script
2.3.1.2	Save the content to the current page
2.3.1.3	If No, don’t apply and script conversion
2.4	Repeat Step 2.2 , apply conversion and save the content whenever required
3	Migrate the space and pages to confluence cloud

WBS

Sl No#	Task	Description	Estimated Effort( Hours)

Sl No#	Task	Description	Estimated Effort( Hours)
1	Development of script
1.1	Setup the dev environment		2
1.2	develop logic to read the list of all the spaces and pages	Logic to read all the pages using REST API	16
1.3	develop logic to read each page and extract the content using REST API		8
1.4	develop logic to check/verify if the conversion script can be applied		8
1.5	develop logic to apply the conversion script to the page		8
1.6	Develop logic to save the converted page content		8
1.7	Develop log to capture the error with report	Log all the errors and generate conversion report	16
2	Testing
		Test the scripts with 5 space	16
			82
3	Script execution	Space 40 and approx. 360 pages Assuming 1 page conversion time : 5 mins *For 360 pages = 3600 5 = 1800 mins = 30 hours**	58
4	Sanity check
4.1	Fix the Based on error and report		16
4.2	Run the script independently to check if any Brikit page left over		8
			24

Total estimated development effort = Development of script + Testing = 82 Hrs

Total Script Execution = Script execution + Sanity check based = 58 Hrs + 24 Hrs = 82 Hrs

Total estimated effort = 164 Hrs

Note: If two resources work on this project, it will take 15 days approx. to complete the conversion

4. Risks & Assumption

Risks

Risk	Priority	Mitigation

Risk	Priority	Mitigation
Access to confluence server	High	Need access to the server before we start the development for all the developers
Access to rest API	High	Need access to the REST APIs
Script execution time	Medium	As per our assumption , each page execution may take 5 mins. Any delay from server/APIs/Network may affect the timelines

Assumptions

Team has access to all the spaces required for the migration
Team has access to server to develop and deploy the script
Script deployed to Empyra’s data center
Business user/power user available for any clarification
Scheduled calls/meetings with power user for any discussions
Weekly status call with business to share the project progress
Post migration, format of the text from confluence server and confluence cloud will be different. We can’t retain the formats.

References

Page links : https://rakutencartera-sandbox-771.atlassian.net/wiki/spaces/ProdMgt/overview , https://rakutencartera-sandbox-771.atlassian.net/wiki/spaces/PM/overview

Atlassian Apps - Internal

POC || Migration ( Confluence Server to Confluence Cloud)