Since approximately 03:11 UTC on 2026-08-28, every CloudKit private-database write from my iCloud account fails. The app had been syncing normally for weeks before this moment.
Failure signature:
Every record save returns HTTP 500 with an empty response body (Server: AppleHttpServer, via icloud-xrail). Client-side error: CKErrorDomain 15 (Server Rejected Request) with underlying CKInternalErrorDomain: 2001. Reads succeed. CKContainer.accountStatus reports available. Authentication is fine. Only writes fail. Reproduces identically via NSPersistentCloudKitContainer mirroring and via a raw CKModifyRecordsOperation probe. Device: iPhone 15 Pro Max, iOS 17.2.1. Development environment, private database.
Evidence this is account-scoped server state, not app configuration:
It reproduces identically in two containers: the original production container and a freshly created container set up after the failures began. A brand-new container failing the same way rules out container-specific corruption or schema issues. Entitlements, embedded provisioning profile, and App ID iCloud capability have all been verified correct (codesign inspection and App Store Connect API). No code or configuration change coincided with the onset. The failure follows my Apple ID across probes over many hours. Every write 500s, without exception.
Correlating event: the legal name on my Apple Account was changed at account.apple.com the same morning, shortly before the failures began. I cannot prove causation, but the timing is exact, and the account-scoped, writes-only signature is consistent with a wedge in the account's server-side state (PCS / user-record layer) introduced by the identity change. I request a check of this account's CloudKit/PCS state.
Sample failing request UUIDs (x-apple-request-uuid) for locating these in server logs:
244DA811-E4AA-4771-88F0-7C092EFD8DEF, 2026-08-28 ~09:57 UTC 1FD544D6-6ED2-4181-8976-B4F4C35830D9 E55EFF8F-FC32-48BD-9165-F4EA263F130A
Many more available on request. The app logs every attempt.
Related reports: this appears closely related to thread 843819 (same week, same signature: private-db writes CKError 15 / HTTP 500, fresh container also affected, Console works), and possibly to the recent cluster of private-database reports in threads 838743, 840248, and 839650. Please correlate rather than triage in isolation.
Impact: this blocks shipping a family-sharing feature built on CKShare, the purchase-deciding feature of the app. All development on sync is stopped.
I can supply on request: a CloudKit sysdiagnose captured during a failing save (per TN3163), additional request UUIDs with timestamps, and the exact local timeline of the account name change.