• Cloudflare R2 returned elevated 503 errors in Eastern North America from 21:10 UTC on 4 September until early on 5 September
  • Applications that depend on R2 can surface storage errors to users even when the incident is limited to one product and region

The fact

Cloudflare marked an R2 object-storage incident in Eastern North America resolved at 00:42 UTC on 5 September. The incident began at 21:10 UTC on 4 September, when customers in the region experienced an elevated number of HTTP 503 errors. Cloudflare's status history says engineers identified the cause, implemented a fix and monitored the service before closing the incident.

The incident was reported for R2 in Eastern North America rather than as a global Cloudflare outage. R2 is Cloudflare's object-storage service, used to hold files, assets and other application data. An application that depends on objects stored through the affected service can therefore return errors to users when access to that storage is disrupted.

The assessment

The incident was limited enough to be described by product and region, but applications using that R2 service could still fail when they tried to retrieve stored data. For an application team, that makes the location of the failure important: an R2 503 points to a storage request that could not be completed, rather than evidence that Cloudflare's entire platform was unavailable.

The next step in diagnosis depends on knowing which parts of an application rely on R2 in the affected region. A service that needs those objects for a critical function may feel the incident immediately, while another may continue operating with only part of its content unavailable.

The supplied reporting does not establish customer-level downtime or data loss, so the impact should not be extended beyond the elevated errors Cloudflare reported. For BTW readers, the resilience decision therefore belongs at workload level. Teams need to know which application functions depend on the storage service before deciding how much protection that dependency requires.

What to watch

Watch for any post-incident explanation from Cloudflare and for further R2 incidents in Eastern North America. More detail on the cause and recovery would clarify the failure boundary and whether the same regional pattern is recurring. Customer reports of access problems continuing after Cloudflare's resolution time would also help establish whether recovery extended beyond the provider's status update.