Way around Azure DevOps performance throttling for Work item migration

Question

I am using the Azure DevOps Migration Tools by Naked Agility and has got them running fine. The work items are getting migrated but in just a few minutes we hit the threshold for rate limits. I am not a developer myself so I'm not sure where to start or if there is a way to go around these limits. We are migrating around 9000 work items and got about half way in 4 days and then the program crashed. So does anyone have tips for how to speed things up?

I restarted the migration now and it skips the ones already there but it will still take days for it to finish.

Answer 1

There is no way to curcomvent the rate limits and they are there to protect other users of the service. If you are migrating a lot of work items then you will definatly hit them, and things will run a little slower.

9000 work items is not a lot, so I am converend that it is tking so long. I recommnned running the migration.exe from a VM in the same data center as the target server.

Answer 2

So I tried with a VM in Azure instead, set up exactly like in the video overview. Copied my migrator folder from the local VM and restarted the migration. After 10 minutes I got a mail from DevOps saying my requests are being delayed:

On 9/8/2020 at 11:01 AM (W. Europe Daylight Time) we detected that your resource consumption on https://dev.azure.com/orgname/ exceeded one of our limits. To maintain service availability for other users, we began delaying some of your requests. Your requests will continue being delayed until your resource consumption returns below our limits.

I have some more thoughts about what can cause this.

• I only have a guest account, since it is a customer env, in the DevOps tenant, but it is added as project collection administrator.
• The DevOps tenant is not in my home Azure tenant.

Now migration has come to 45 items in about 30 minutes and they have all been skipped since it is restarted. The average time is around 38 seconds so I expect it to crash when we hit over 60 seconds.

And by the way, the latest build (10.0) of the tool does not seem to work. Crashed on NodeStructure config although it was set to false.