# Cloudflare Logged 13 Incidents in 8 Days, One Still Unresolved

Page: https://hirantha.me/blog/cloudflare-logged-13-incidents-in-8-days-one-still-unresolved
Published: 2026-08-26
Written by: hirantha.me Newsroom (AI-generated). An AI model wrote this from the sources listed below; how it is made: https://hirantha.me/about/newsroom
Topics: cloudflare, r2 storage, cloud reliability, site reliability, cloud outages

> Cloudflare's status page recorded 13 incidents in eight days this month, and a customer's reported data loss remains unconfirmed.

## Thirteen incidents, eight days

Between August 7 and August 14, 2026, Cloudflare's public status page recorded thirteen separate incidents. That works out to more than one and a half a day on average, though the real shape of it is bursty: three quiet days in the middle of the window, then four incidents apiece on August 13 and August 14 alone.

Twelve of the thirteen carried a minor-severity label. The one exception, on August 12, was a major-rated disruption to outbound email delivery caused by one of Cloudflare's sending domains landing on a Spamhaus blocklist.

## What actually broke

The cluster opened on August 7 with a write-availability problem in a small number of R2 object storage buckets in Cloudflare's Eastern North America region, running from 14:52 to 17:02 UTC; Cloudflare says it identified the cause and mitigated further impact by 22:54 UTC that day. The following day, Istanbul lost network performance after Cloudflare reported losing a dark fiber link and had to coordinate restoration with outside vendors.

August 13 was the busiest single day: elevated errors on Workers KV, a key-value storage product; 503 errors on Magic Transit, Cloudflare's network-layer DDoS protection service; authentication failures on the Model Context Protocol (MCP) Server Portal, the interface Cloudflare offers for managing AI agent tool connections; and problems with Workers AI, its inference platform. August 14 brought four more incidents, including a Durable Objects and Workflows availability drop plus regional network errors spanning the eastern US, the Middle East and Southeast Asia, and Mexico. Between those two peak days, customers as far apart as Kuwait, Bangkok, Jakarta, and Dammam saw regional spikes in server-side errors, underscoring how a single provider's infrastructure problems can surface as unrelated-looking symptoms in completely different parts of the world at the same time.

## The data question nobody has answered

Days after the August 7 R2 incident, a customer posted on Cloudflare's community forum reporting that roughly 67 gigabytes of data in one of their R2 buckets had not been restored. As of that report, Cloudflare had not published a formal postmortem confirming or ruling out permanent data loss tied to the incident. R2 is Cloudflare's S3-compatible object storage product, marketed in part on the promise of no egress fees, and it has become a common backing store for everything from application backups to media libraries, which is exactly the kind of data where a partial, unacknowledged loss is hardest to notice until someone goes looking for a specific file. That silence matters more than any single minor-severity label: a status page marking an issue "resolved" tells you the service came back, not that every byte that was in flight during the outage came back with it.

## It didn't stop on the 14th

Cloudflare's own incident history shows the company's rough patch didn't end when the eight-day cluster did. Later in August, the status page logged a Durable Objects and D1 database degradation across Hong Kong and Singapore lasting roughly four and a half hours on August 24, followed by an increased-latency issue on its CDN and cache layer that ran for about eighteen hours into August 25. Both were labeled minor, consistent with the rest of the month, but the pattern of frequent small failures across different products and regions has continued rather than resolved into a single clean root cause.

## What it means in practice

None of these incidents individually looks like the kind of headline-grabbing, hours-long global outage that makes international news. That's precisely the risk worth paying attention to: a service that fronts a large share of the modern web having a stretch of frequent small failures across storage, compute, and networking products is a different kind of warning sign than one big outage, because it's easy to miss if you're only watching for red banners. Teams building on Cloudflare's R2, Workers KV, or Durable Objects should treat this month as a prompt to check their own backup and replication strategy for anything stored there, set up independent monitoring rather than relying solely on Cloudflare's own status page, and follow up on whether Cloudflare ultimately publishes a postmortem addressing the unresolved data report before assuming the incident is fully closed.

## Questions

**What happened to Cloudflare in August 2026?**
Cloudflare's status page logged 13 separate incidents between August 7 and August 14, touching R2 storage, Workers KV, Durable Objects, Workers AI, Magic Transit, email delivery, and regional network performance, and further minor incidents continued later in the month.

**Was Cloudflare completely down?**
No. Twelve of the thirteen incidents were rated minor and affected specific services or regions rather than the whole network; only the August 12 email delivery disruption was rated major.

**Was any customer data actually lost?**
A customer reported on Cloudflare's community forum that about 67 gigabytes of R2 data was not restored after the August 7 incident, but Cloudflare had not published a postmortem confirming or denying permanent loss.

**What should businesses that rely on Cloudflare do?**
Maintain independent backups or replication for critical data stored in Cloudflare products, monitor status directly rather than assuming no news is good news, and watch for a formal postmortem on the unresolved data report.

## Sources

- [Cloudflare Logs 13 Outages in 8 Days as R2 Falters](https://shattered.io/cloudflare-outage-august-2026/), Shattered.io
- [Cloudflare Outage: Latest Status and What Happened](https://www.itechguides.com/cloudfare-experienced-a-global-network-outage-heres-the-latest-updates/), iTech Guides
- [Incident History](https://www.cloudflarestatus.com/history), Cloudflare Status
- [R2 data loss after ENAM incident Aug 7-8 - 67GB not restored](https://community.cloudflare.com/t/r2-data-loss-after-enam-incident-aug-7-8-67gb-not-restored-bucket-comfy-storage/947037), Cloudflare Community