In 2016, there were a total of 525 cloud outage incidents spread across 15 of the most popular global cloud service providers.
17% of outages lasted for less than 1 hour; 51% lasted between 1 and 5 hours, and the remaining 32% lasted anywhere between 5 and 72 hours.
Cloud Computing:
Power & Frailty
For all its operational, scale, and cost advantages, cloud computing is an (un)surprisingly fragile technology.
Introduction
In 2016, there were more than 500 publicly recorded instances of cloud failures across more than 15 global cloud service providers. Of these, roughly 17% lasted less than 1 hour, 51% between 1 and 5 hours of downtime, and the remaining 32% lasted anywhere between 5 and 72 hours.
The financial ramifications can hardly be understated. 2016 saw destruction of more than $3.8 billion of business value. When the cloud services that businesses count on fail, operations are disrupted and real financial pain manifest as: forgone revenue, lost transactions, breach of end-customer trust, refunds and claims, related regulatory fines, and even lawsuits in some extreme cases. So what are the lessons here?
-
No infrastructure - whether it's hosted on the cloud or on a local data center - is failproof.
-
Operational recovery from cloud service outages is not instantaneous and businesses are often left at the mercy of the service providers.
-
High-availability cloud architecture is expensive, complex, and most businesses don't actually implement it optimally.
-
The current state of cloud computing and financial systems leave a major void that cannot be filled using traditional risk-management solutions.
Jump to...




17%
33%
50%
% SMB Workloads in Cloud
Survey Sample: 517
% Enterprise Workloads in Cloud
Survey Sample: 458

25%
43%
32%
% Workloads in Cloud
Survey Sample: 1,002

21%
38%
41%
Non Cloud
Public Cloud
Private Cloud
Data courtesy of RightScale State of the Cloud Report 2017
17%
51%
32%
*Oracle Cloud, HP, Alibaba Cloud, CenturyLink, Linode Cloud, Digital Ocean, City Cloud, Faction Cloud, GoDaddy Cloud, Elastic Hosts.


2016 Cloud Outage Distribution
# Incidents
Downtime Hours
*
Amazon Web Services
Compute: EC2, Elastic Bean Stalk, VPC, Lambda,
Auto-Scaling
Storage: AWS S3, EBS, Glacier, Elastic File System
Database: AWS RDS, Dynamo DB, Simple DB, Aurora, Elastic Cache, RedShift
Microsoft Azure
Compute: Azure Virtual Machine, App Service, Functions, Azure Container Service
Storage: Azure Storage, Data Lake Store, StorSimple
Database: Azure SQL Database, MYSQL, PostgreSQL, DW
Google Cloud Platform
Compute: Google Compute Engine, App Engine, Container Engine
Storage: Google Storage, Persistent Disk
Database: GCP BigQuery, SQL, Big Query, Dataflow
Rackspace
Compute: Rackspace Cloud Servers, Cloud Load Balancers
Storage: Rackspace Storage
Database: Rackspace Big Data, Database
IBM Softlayer
Compute: IBM Virtual Servers
Storage: IBM Storage
Database: IBM Big Data
Cloud Outage Category
