# Cloudflare: Cache Purging Delays in Amsterdam · ORBIT Outage Dissector

- Provider: Cloudflare · state: resolved · duration: 1h 41m
- CHYP Magnitude: **M4.4 Squall** (severity 0.45, breadth 0.37, duration 0.5, reach 0.54, corroboration 0.25)
- Content SHA-256: 10b0ab010e14983f45e1aa5fe8c2a2d53a1e313ecc11dcb008cd76cbb3b77181

## Story

- **What happened:** Cloudflare reported “Cloudflare: Cache Purging Delays in Amsterdam” starting Nov 20, 10:01 AM PT.
- **What broke:** Ahmedabad, India - (AMD), Albuquerque, United States - (ABQ), Arica, Chile - (ARI), Annaba, Algeria - (AAE) and 2 more.
- **Who was hit:** 14 catalogued services run on Cloudflare (inferred); 2 of them show an official problem in the same window (coincidence, not causation). Hit: 1 named location (Amsterdam (AMS)).
- **How it recovered:** Resolved after 1h 41m.

## Failure-mode fingerprint (automated keyword read (hypothesis, not root cause))

- Control-plane coupling: 90% (“purging”, “purge”, “dashboard”)
- Config propagation: 60% (“updates”, “changes”)
- Dependency cascade: 30% (2 dependents also unwell)

## Timeline (PT)

### Official status
- Nov 20, 10:02 AM PT · Cloudflare: Cache Purging Delays in Amsterdam
- Nov 20, 11:42 AM PT · Cloudflare: Cache Purging Delays in Amsterdam

### Network & routing
- Nov 20, 11:42 AM PT · AS13335 (Cloudflare) BGP update churn +0.2σ vs 24h (https://www.chyp.ai/incident/event-nw-7ef1b996828d7b7e)

### Early signals

### News

### Coincidence (KEV)

## Historical twins

- 60% · Cloudflare 2022-06-21: Network configuration change took 19 data centers offline (75m) https://www.chyp.ai/dissect/m/cf-mcp-bgp-2022-06-21/
- 57% · Cloudflare 2025-11-18: Bot Management feature file doubled and crashed traffic-routing software (3h 10m to core traffic; 5h 46m to all systems) https://www.chyp.ai/dissect/m/cf-bot-feature-file-2025-11-18/
- 56% · Cloudflare 2020-07-17: Backbone router configuration error overwhelmed Atlanta (27m) https://www.chyp.ai/dissect/m/cf-backbone-2020-07-17/

## Cloudy_Day take (automated)

- The public wording leans control plane (APIs, dashboards, purges, provisioning). If that holds, it's the management layer on the critical path again. Worth asking.
- 2 services that sit on Cloudflare are unwell at the same time. That's a lead to check, not proof of a cascade.

## Sources

- Cloudflare official status page: https://www.cloudflarestatus.com/incidents/xsgbfncq7of7
- NetWeather evidence page: https://www.chyp.ai/incident/event-nw-dcf64fde3af020e5

## Limits

- Provider-stated: this repeats what the vendor's public status page said. It is not an independent root-cause verification.
- Temporal overlap with other events is coincidence, not causation, unless a provider states a causal link.
- NetWeather is independent research, not an official status source. Confirm with the provider.
- Correlation is not causation.

> Cite: ORBIT by Cloudy_Day, Outage Dissector, https://www.chyp.ai/dissect/event-nw-dcf64fde3af020e5/
> Automated data from ORBIT (Cloudy_Day Security Suite). Independent research, not affiliated with any government agency, vendor or data provider. Not an official status source.
