Mass import Gliffy diagrams to draw.io in Confluence Server
The mass Gliffy import converts all Gliffy diagrams on all pages (or selected spaces) to draw.io diagrams. Both the current Gliffy JSON format and legacy XML formats are supported.
Test the mass import on a mirrored test instance first.
Confluence must be re-indexed prior to performing the mass import so all diagrams are found.
The Gliffy app must be installed, but it’s recommended to remove the license to make Gliffy diagrams effectively read-only during the mass import.
Run the Gliffy mass import feature
As an administrator, install the draw.io app.
Go to the Atlassian Marketplace heading in the left panel of your instance’s administration area, then click on Find new apps.
Search for draw.io and install the draw.io Diagrams for Confluence app.
Once it has been installed, under the draw.io add-on heading in left panel of your instance’s administration area, then click on Configuration and select the Gliffy Import tab.
To limit the spaces you import, enter a filter into the Filter by spaces field. Use this to test the import on a small demonstration space. Then click on Start Import.
Note: The full import may take some time depending on the number of Gliffy diagrams there are. A text report of the import is created, which you can download and use to inspect the results once the import is finished.
Additional configuration options
You can add the following configuration options by adding URL parameters to the draw.io Configuration page’s URL.
Configure a comment to be applied to the converted diagrams by adding the pageVersionComment=YourURIEncodedString URL parameter.
Ensure that the date of the last modification made to each page isn’t changed by adding the keepDates=1 URL parameter.
For example, replace BASE_URL with the URL of your Confluence Server instance:
It is better to perform the conversion outside of your company’s normal usage hours of the Confluence instance. You can do the import during normal working hours, but it’s best to carry out the whole operation whilst the minimum number of changes are being made to the relevant pages.
We recommend that you remove the Gliffy license so that users do not start editing Gliffy diagrams while the process is running. Without a license, the Gliffy diagrams are effectively read-only. Please communicate to your users that you are doing this and when you expect to start/finish the process.
draw.io automatically creates a new page revision of every page that contains a Gliffy diagram. You can always revert pages to their previous versions if needed as no Gliffy data is deleted.
When you buy an app like Gliffy, you get a perpetual license to any version if the release date is prior to the end of your maintenance period. These versions will continue to work beyond this expiration date, allowing you both to view and edit Gliffy diagrams.
Most Confluence administrators prefer to unambiguously recommend users diagram with draw.io and so uninstall the Gliffy app. If Gliffy is already uninstalled, the mass Gliffy to draw.io import will still work, as the diagram files are still attached to their pages.
Note that we don’t recommend you uninstall Gliffy until after the mass import is completed and you have checked the result.
An administrator who has full write permissions across all spaces in your Confluence instance should run the mass import. If no such user exists, you will need to temporarily elevate an administrator to have those permissions for the duration of the import.
Make sure you always test the mass Gliffy import on a test instance with your production data first. If you encounter any issues during this mass import, please email sup[email protected] sending Gliffy Import Report text file as an attachment to your email.
Workaround for the CONFSERVER-59196 bug
Confluence version 7.0.0 and later has a bug which requires vendors to disable bulk operations such as the mass Gliffy import.
As a workaround, update the draw.io app to version 9.2.7 or later and add the ignoreBulkOpsWarning=1 URL parameter to the draw.io Configuration page’s URL. For example:
This workaround will simply re-try a failed operation, there is no risk of data loss involved. However, we cannot guarantee that the workaround will be 100% successful.