SYSTEM Cited by 1 source
Cloudflare Cloud Observatory¶
Cloudflare Cloud Observatory is a Cloudflare Radar surface at radar.cloudflare.com/cloud-observatory that publishes per-cloud-provider, per-region connection health metrics derived from Cloudflare's edge attempting to retrieve uncacheable or cache-expired content from origin servers hosted in each cloud region.
What it exposes¶
For each hyperscaler (Amazon, Google, Microsoft, etc.) and each region within it (e.g. me-central-1, me-south-1), the observatory reports:
- Connection failure rate — the rate at which Cloudflare's edge fails to establish a successful connection to an origin in that region when retrieving uncacheable content. See concepts/connection-failure-rate for the metric definition.
- Trend lines — historical per-region time series so sudden degradations are visually obvious against baseline.
- Per-event URLs — shareable deep-links with date ranges
(
?dateStart=…&dateEnd=…#connection-metrics) that pin the graph to an incident window, used directly in Cloudflare's quarterly disruption reviews as evidence.
Why Cloudflare has this view¶
Cloudflare acts as a reverse proxy for a large share of the Web. For every cache miss on Cloudflare-proxied sites whose origin is in a cloud region, the edge opens a fresh TCP + TLS connection to the origin. Success rates aggregate into a third-party uptime signal for that region — independent of the cloud provider's own status dashboard. Where AWS / GCP / Azure status pages reflect the provider's internal view of whether services are operational, the Cloud Observatory reflects whether external callers can actually connect.
This matters most when:
- A provider's status dashboard lags the real outage.
- The fault is in the connectivity path, not the service.
- Physical damage affects the facility but not the control plane.
- A workload needs to decide quickly whether to fail over to another region (the basis for patterns/cloud-region-migration-during-conflict).
Seen in¶
- sources/2026-04-28-cloudflare-q1-2026-internet-disruption-summary — canonical instance of Cloud Observatory as third-party hyperscaler outage forensics. After drone strikes on Amazon data centers in the UAE and Bahrain on March 1-2, 2026, the observatory's me-central-1 and me-south-1 views showed "elevated connection failure rates... beginning March 1-2 and remaining higher for multiple days." The review explicitly defines the metric: "Connection failures occur when Cloudflare fails to successfully connect to an origin server when attempting to retrieve uncacheable content, or content not in/expired from cache. These graphs illustrate the increased rate of failures experienced when attempting to connect to servers in these impacted regions." First wiki instance of Cloud Observatory as the visibility layer for kinetic attacks on cloud infrastructure — independent of and complementary to the AWS Health Dashboard.