I tried to reproduce this on the current build and could not.
Not run against a live Azure DevOps org: no worker has an Azure DevOps organisation or az sign-in. Reproduced instead with a unit test on origin/main 6f21b0f02a in packages/source-control-azure-devops/src/server/AzureDevOpsPullRequestProvider.test.ts. The test mocks AzureDevOpsPullRequestCli with 200 changed files, and each readItemContent sleeps 2.5s (the reporter's measured az startup time). It forks provider.getDiff and moves the TestClock forward 60s (the client's DEFAULT_PULL_REQUEST_DIFF_TIMEOUT_MS). The first page is still pending, so the test fails. As a check, the same test passes when the clock moves 130s and fails at 120s, so the first page takes about 125s: 50 sequential batches of 4 files at 2.5s each. This is the mechanism behind the reporter's trace, where the request was interrupted at 60,005 ms with readItemContent calls still running.
Closing as not planned. If it still happens for you, reply with the exact steps and the version you are on and it will be looked at again.