To read more about Azure DevOps data redundancy, see : https://learn.microsoft.com/en-us/azure/devops/organizations/security/data-protection?view=azure-devops#data-redundancy
To protect data during hardware or service failures, Azure Storage geo-replicates customer data between two regions in the same geography. For example, Azure can geo-replicate data between North and West Europe or between North and South United States. For Azure Blob Storage, customer data gets replicated three times within a single region, and is replicated asynchronously to a second region in the same geography. As such, Azure always maintains the equivalent of six copies of your data. This enables you to fail over to a separate region if there's a major outage or disaster, while also having local redundancy for hardware failures within a region. For Azure SQL Database storage, daily backups are maintained offsite if there's a regional disaster. Microsoft also takes point-in-time backups of both the blobs in Azure Blob Storage, and the databases in Azure SQL Database. There's a separate copy of all blobs, and changes are appended to each storage account. Because this data is immutable, you don't need to rewrite any existing storage as part of the backup procedures. Backups are a standard part of Azure SQL Database, and Azure DevOps makes use of this. We maintain 28 days' worth of your data
This will be as specific as you will be able to get.
Where your repository contains valuable IP, you can of course orchestrate your own backups. For this i'd suggest looking at the Azure DevOps CLI.