Zephyr Squad migration duration

[Ciocirlan_ Costel-Iulian]

Hello everyone,

We have recently started doing test migration from Jira Software DC to Jira Cloud, and you already know, we also need to bring along the add-ons that our customers are using in DC.

One of the add-ons that we will move with in the Cloud is the Zephyr Squad Test management add-on.
We have engaged Atlassian and SmartBear support on this, the overall performance of project data migrations using JCMA as improved but I cannot say that this is the case for Zephyr.

We have several projects using the add-on, most of them with not a lot of data, but there is 1 project which has a big number of Zephyr data which needs to be migrated.
For anyone who is aware of the http://jira_server_baseurl/rest/zapi/latest/migrations URL, this shows the progress of the Zephyr data migration with the number of Zephyr attachments that have been successfully uploaded.

To have an idea on how big the 1 project is, the migration ran for 3 weeks with Zephyr data migration at 0% and we got to ATTACHMENTS_PART_58 till we decided to stop it.
Migrating a smaller project that also took quite a lot of time we only got to ATTACHMENTS_PART_1 - so indeed there is a big difference of sizing between the two projects.

However, even this small project took a lot of time compared to the project data size: pre-checks took 2 minutes, project data migration took 10 minutes and the Zephyr data migration took 1 hours & 42 minutes.

The question that I have is: from your experience, do you have any advice on how to boost the performance of the Zephyr Squad migration?

I want to mention again that we have been in discussions with Atlassian and SmartBear and followed their advice and recommendations already, but we are at the point where the only remaining advice that we have from SmarBear is to remove older Zephyr data from the project and audit history, but without giving us at least an approximation on how much "older data" should we remove in order to have this migration complete.
And taking into consideration that even for the small project took 1 hour & 42 minutes for the Zephyr migration, I don't feel comfortable going to our customers asking "Hey, we need to delete some data because otherwise we will be stuck 0% for more than 3 weeks, but I am not aware of how many data".

Thank you in advance for any support!
Costel

[Carl Storey]

hi @Ciocirlan_ Costel-Iulian we currently have a very similar issue trying to migrate a large amount of Zephyr data from DC to Cloud. Did you manage to overcome this in the end? It would be really useful to know.

Thanks

Carl

[Ciocirlan_ Costel-Iulian]

Hi @Carl Storey ,

What we ultimately did was archive Test Executions that were older than 1 year Archiving Test Executions | Zephyr Essential DC Documentation

You should also make sure that audit Zephyr data is not included in the migration (if you of course don't need it). I don’t remember the exact location (we did this back in 2024), but there was a toggle somewhere in the Zephyr settings under Manage Apps, if I recall correctly.

Additionally, you can archive older test cases (Test issues) as well.

Of course, this is something we aligned on with our end users beforehand and got their agreement before proceeding.

Beyond archiving, we kept close contact with SmartBear. They were aware of our production migration timeline and also supported test migrations when needed. They can actually reserve a migration seat for you, so your migration doesn’t get stuck in a queue.

I strongly recommend reaching out to @Luzia Mendes [SmartBear]  for this, she can provide more details. Also, I’d assume things have improved since 2024, so they might have new suggestions or optimizations to share.
This step significantly improved the performance and speed of our Zephyr data migration.

What did we do with the archived Zephyr data?
In our case, there was one large project containing ~99% of the old data. Before starting the archiving process, we mirrored our production environment.

After completing the production migration, we ran a secondary migration from the mirrored environment (containing the archived Zephyr data) to a separate Cloud site within our Atlassian Cloud organization. Access to this site was limited to org admins and a few project admins, so we didn’t mind that this migration took over a week via JCMA.

We actually decommissioned that site recently after 2 years - it turned out not to be necessary anymore. However, if you want your users to still have access to historical data “just in case,” this approach can work well.

Quick recap of our approach:

1. Contact SmartBear to reserve a migration seat and inform them about your migration (provide your Cloud URL).

2. Mirror your production environment.

3. Confirm with users whether older data can be archived.

Before starting the archiving:
       4. In Zephyr settings:

• Clear cache

• Clear permission cache

• Re-index Zephyr project metadata (all projects)

• Re-index executions (all projects)

5. Perform a full Jira re-index

6. Disable the setting to migrate audit Zephyr data (if you don't need the data)

7. Review Estimate Your Migration Estimate Your Migration | Zephyr Essential DC Documentation

Archiving steps:
       8. Archive test executions via bulk archive Archiving Test Executions | Zephyr Essential DC Documentation
       9. Archive test issues using JQL → bulk archive

After archiving:
       10. In Zephyr settings:

• Clear cache

• Clear permission cache

• Re-index Zephyr project metadata (all projects)

• Re-index executions (all projects)

11. Perform another full Jira re-index

12. Review Estimate Your Migration Estimate Your Migration | Zephyr Essential DC Documentation
13. Run a test migration a see if all the above improved your migration performance/speed
    
    

Hope this helps 🙂

Costel

[Carl Storey]

hi @Ciocirlan_ Costel-Iulian 

Thanks so much for this info! I will share it with our project team. We actually have a call with @Luzia Mendes [SmartBear] about this later on so hopefully we can come up with a similar solution for our migration issues.

All the best,

Carl 

[Ciocirlan_ Costel-Iulian]

Up!

[Luzia Mendes [SmartBear]]

Hi Costel, 
Glad to see you here in the Community. 

This project of yours is indeed an "interesting" challenge for us.

I will need to discuss the case internally with my group to see if any creative ideas might help us. So let's hope that the community helps us here :) 

Cheers,
Luzia Mendes
Zephyr Migrations Manager