The format of the Teams compliance records generated for personal and group chats and stored in Exchange Online mailboxes is changing. Microsoft is removing a bunch of unnecessary attributes from the records to reduce the processing load on the service to retrieve the attributes from Azure AD. The change is unlikely to affect most tenants. Compliance records for older chats are not affected.
Because it sits on top of so many Microsoft 365 components, Teams is easily the hardest Office 365 workload to backup. You can try to backup Teams by copying its compliance records stored in Exchange Online, but that’s only a partial (and bad) solution that utterly fails to take the full spectrum of Teams data into account.
For compliance purposes, the Microsoft 365 substrate captures copies of Teams messages in Exchange Online mailboxes. The compliance records are indexed and discoverable, which means that they can be found by content searches. However, Teams compliance records are imperfect copies of the real data, which is a fact that seems to have escaped many people.
Teams does a good job of storing compliance records in Exchange Online mailboxes so that the data is available for Office 365 eDiscovery. But the number of records can impact the mailbox quotas of frontline workers, especially if they send graphics in personal and group chats. Here’s some PowerShell to help discover how much mailbox quota is being absorbed by compliance records.