Retry amplification
Failed or slow calls produce more client or peer retries, which increase load, which produce more failures — a retry storm, thundering herd, or metastable collapse.
CHYP.ai · ORBIT // Cloudy_Day Security Suite
ORBIT · CLOUDY_DAY · FLAGSHIP
Every outage, taken apart. Pick any live or historic incident and get the reconstructed timeline, blast radius, blast map, failure-mode fingerprint, evidence matrix, historical twins and one honest magnitude, from public sources only.
● live right now (pulsing) · ● famous meltdowns, coloured by primary failure mode. Up is worse, right is longer. Click any dot to dissect it.
Cloudflare: 19 components partial outage · AS
open 2d · AS · 14 dependents
Dissect →Cloudflare: 6 components partial outage · AF
open 2d · AF · 14 dependents
Dissect →Cloudflare: 4 components partial outage · SA
open 2d · SA · 14 dependents
Dissect →Cloudflare: 3 components partial outage · NA
open 2d · NA · 14 dependents
Dissect →AWS Bahrain: Multiple services — Region Availability
open 2d · ME · 136 dependents
Dissect →AWS UAE: Multiple services — Region Availability
open 2d · ME · 136 dependents
Dissect →Cloudflare: 2 components partial outage · EU
open 2d · EU · 14 dependents
Dissect →Cloudflare: 2 components partial outage · OC
open 2d · OC · 14 dependents
Dissect →[Network-Wide] Zetachain Testnet Stall
open 42h 21m
Dissect →USPS International Label Creation Impacted for Select European Destinations
open 20h 36m
Dissect →Cloudflare: 2 components degraded performance
open 2d · 14 dependents
Dissect →INC20000369
open 5h 36m
Dissect →Exchange / Microsoft 365: servers unexpectedly unavailable, core process overwhelmed
CapacityCascade
Ongoing as of 3:19 PM PT (no official full-restore time in the sourced notes)
Network-control-plane configuration update dropped GCVE private traffic
Config pushControl plane
10h 40m
WAF change for a React vulnerability hit a latent Lua bug
Config push
25m
Bot Management feature file doubled and crashed traffic-routing software
Config pushCapacityCascade
3h 10m to core traffic; 5h 46m to all systems
Front Door: incompatible metadata across two control-plane versions
Control planeConfig pushCascade
8h 24m
DynamoDB DNS automation wrote an empty record it could not repair
Control planeCascade
14h 32m overall (three distinct impact periods)
Front Door: manual metadata cleanup bypassed protection
Config pushControl plane
8h 10m (availability restored 12:50 UTC)
Service Control crash loop after a blank quota-policy field went global
Control planeConfig pushCapacity
7h 27m
ARM control plane blocked by a global Cosmos DB misconfiguration
Config pushControl plane
1h 38m
Front Door: network control-plane failure plus latent routing configuration
Control planeConfig push
7h 58m (majority by 13:58 UTC)
Kinesis cell-management misidentified healthy hosts during a deployment
Config pushCapacityCascade
6h 52m
Sensor content-configuration update caused a widespread Windows outage
Config push
duration not stated
New DDoS rule exposed a latent rate-limiting bug
Config push
114m
Lambda cell crossed a capacity threshold and hit a latent defect
CapacityControl planeCascade
3h 48m to full recovery
Network configuration change took 19 data centers offline
Config push
75m
Internal-network surge, retries, and control-plane API failures
RetriesControl planeCapacity
6h 52m to network-device recovery
Valid customer configuration triggered a latent global bug
Config push
49m to 95% of the network
Automated quota system shrank User ID Service capacity
CapacityConfig pushControl plane
47m
Kinesis front-end fleet exceeded OS thread limit
CapacityControl planeCascade
17h 8m
Backbone router configuration error overwhelmed Atlanta
Config pushCapacity
27m
IAM configuration change propagated globally past canary
Config pushControl plane
50m to majority mitigation
From the Control-Plane Meltdown Index: vendor post-incident reports only, quotes verbatim. Open the index (CSV, JSON) →
Failed or slow calls produce more client or peer retries, which increase load, which produce more failures — a retry storm, thundering herd, or metastable collapse.
A shared API, metadata, identity, DNS, or management plane sits on the path of many services. When it fails, the blast radius is far larger than the originating component.
A configuration, feature-file, content, or policy update is pushed — globally or past a protection layer — and a latent defect or incompatible metadata takes the data plane down with it.
One impaired foundation (DNS, IAM, streams, a front door) causes a chain of failures in services that depend on it, often with health-check or failover side effects.
A thread limit, quota, health-check eviction, or overwhelmed core process removes capacity; remaining capacity is then overloaded, often accelerating the loss.
One open number from 0 to 10, so a 25-minute global meltdown and a 3-day regional drizzle stop sharing the same headline. Every input is public and shown on each case file.
0+ Tremor 3+ Squall 5+ Storm 7+ Severe storm 8.5+ Meltdown
A ranking aid, not a measurement of harm. It never says why something broke.
Machine-readable: /api/v1/dissect.json · every case file has a JSON and Markdown capsule.