Simulation Testing - Customer Communication

Overview

Jira Software Server and Jira Software Cloud have different architecture and software implementation. Therefore Atlassian wants to ensure excellent user experience is maintained post migrating to the cloud.

How do we achieve seamless user experience?

Jira Software is a versatile product and can be used in many different ways, resulting in a wide range of data shapes, size and traffic. We look at our enterprise customer’s usage profile to gain understanding on how Jira is currently used and simulate that profile using our tools.

We then run load test against that simulation to identify and fix performance anomalies.

What will be captured as part of the customer profile?

Atlassian will not capture any user data. Synthetic data will be generated based on the customer’s data shape. The data shape mainly includes number and sizes of elements. There is no way to infer user generated content from the generated synthetic data.

Who will have access to this simulation?

Access will be internal only and granted to the Performance and scale team for running load test for the purpose of identifying performance anomalies. Access might also be granted to development teams who are working to optimise the user experience for such scale and teams planning and supporting the migration.

All access is inline with the Atlassian’s security and privacy policies.

How will Atlassian capture the customer profile?

There are 3 steps required to be performed by the customer:

  • Backup - The customer will generate a local backup using the standard Jira Backup utility (Option 2).

  • Generate synthetic data - The customer will download a Atlassian provided tool and will run it locally. The tool will generate two files on the local server.

  • Upload - The customer can inspect and upload the synthetic data files to Atlassian’s server.

The entire process duration shouldn’t take more than one day but is depending on the customer’s network connection.

What will happen next?

Atlassian will use the synthetic data to create a Jira cloud instance exhibiting the same profile as the customer. This will ensure reliable simulation testing and prediction of the customer’s user experience.

In case an performance anomaly is identified, development teams will address the issue and verify the issue actually resolved on the simulation and that there are not other underlying issues.

When the test results are all good, the customer can start migrating to the cloud and should expect a smoother transition compared to customers not running the simulation testing.

Will my data be secured?  

The customer data does not leave the customer’s Server/DC site. The backup will be generated locally and can be deleted straight after the customer profile been generated.

The customer profile will be uploaded to our cloud production environment which has all the security controls and mechanisms to prevent unauthorised access to the simulation.

Technical Details

From technical standpoint, what is important to know 

What’s in the customer profile?

The standard Jira backup process creates a zipped file. Once you unzipped the backup file it consist of the following format:

Click here to expand...

JIRA-backup-20230214
├── activeobjects.xml
├── entities.xml
├── data
│   ├── attachments
│   └── avatars
└── logos


The entire Jira database entity data (i.e. Projects, Issues, Comments, Workflows, Permission groups, etc) is stored in these two files activeobjects.xml and entities.xml will be used to create the customer profile.

The downloaded tool for generating the customer profile will inspect the above two files and will generate synthetic data in two different files.

Note that media files like Issue attachments, user avatars and project logos are stored in corresponding directories, but won’t be part of the customer profile.


What will the customer profile look like?


Original Data Generated Data
User

User Profile

Group

Group detail

Project

Last modified on Nov 5, 2023

Was this helpful?

Yes
No
Provide feedback about this article
Powered by Confluence and Scroll Viewport.