Top 7 most catastrophic cloud outages of 2019

artificial intelligence

Table of Contents

The year 2019 was rocked by numerous Cloud outages, reminding the world that the cloud is still far from foolproof and raising doubts about the risks linked to dependence on this new technology … discover the most catastrophic Cloud outages of the year that ends.

Over the past few years, Cloud adoption by businesses and individuals has literally exploded. This is not surprising, since the cloud brings many advantages in terms of flexibility, simplicity and even cost.

By turning to Cloud storage, it is possible to access its files and documents from any device and from anywhere. Likewise, by turning to SaaS cloud services or PaaS, companies no longer need to develop their own infrastructure. They therefore benefit from a simplicity gain while reducing their costs.

However, cloud users also do the choice to entrust the security and safety of their data to their suppliers. By migrating their data and applications to a supplier’s servers, rather than their own data centers, companies are willing to place their trust in them.

However, several breakdowns occurring throughout the year cast doubt on the reliability of these services … discover in this file The 7 Most Disastrous Cloud Outages Of 2019.


In May 2019, deploying a database script on the Pardot Marketing Cloud service from Salesforce caused a serious incident. Ordinary users have been granted higher level permissions.

To prevent employees from stealing sensitive data from their companies, Salesforce had to block many users then block access to other services such as Sales Cloud and Service Cloud.

For more than 20 hours, customers were unable to access Pardot Marketing Cloud. In total, it took 12 days so that other services such as Sales Cloud and Service Cloud are deployed. The entire Salesforce Cloud infrastructure was therefore impacted by a simple script …

Amazon Web Services

Even the world leader in the cloud is not foolproof. In August 2019, a US-EAST-1 Data Center belonging to AWS and located in North Virginia was hit by a power outage. The data center backup generators therefore broke down.

Consequently, 7.5% of EC2 instances and EBS volumes remained temporarily unavailable. Worse still: after the power was restored, Amazon announced that some of the data stored on the damaged hardware could not be recovered.

Some customers have therefore definitely lost valuable information. This incident demonstrates that the cloud is not always synonymous with security, and that a local backup for the most precious data is essential.

Apple iCloud

In July 2019, many Apple iCloud users stayed unable to access the service for several hours. The message “Service Unavailable – DNS Failure” appeared on the screen.

Several Apple services such as the App Store, Apple Music, Apple TV, Apple Books and Apple ID were impacted. Likewise, features like “Find My iPhone” were unavailable during the incident.

According to Apple, this failure is related to a “BGP route flap” problem which has caused significant packet loss for users in North America.

Microsoft Azure

In May 2019, incorrect server name delegation affected Microsoft Azure’s DNS resolution and network connectivity. during More than an hour, Microsoft Office 365, Microsoft Teams or Xbox Live services remained inaccessible.

An hour that can seem like an eternity for the biggest commercial customers of the giant Redmond. Fortunately, DNS records of clients have not been impacted after restoration of services.

Google Cloud

Google Cloud suffered two major outages in 2019. In July, a problem with Cloud Networking and Load Balancing forced Google to separate servers in the US-east1 region from the rest of the world.

This incident caused physical damage to multiple competing fiber bundles serving the region’s network bridges. Many users have experienced significant lag times.

More recently, in November 2019, several Google Cloud Platform services have been impacted by significant issues. This is the case with Cloud Dataflow, Cloud storage, and the Compute Engine. Many products have therefore been affected worldwide.

The Google Cloud APIs were affected on us-east1, us-east4 and southamerican-east1, and some APIs have been affected worldwide. A few days ago, users were faced with a 100% packet loss on approximately 20% of GCP instances in the us-west1-b region for 2.5 hours.


In July 2019, Cloudflare visitors received 502 errors. These errors were caused by a spike in CPU usage on the network.

This spike was caused by failed software deployment. For 30 minutes, the service remained down until the deployment was canceled.

The company had to reassure customers by explaining that it was not a cyber attack. Obviously, a full investigation was carried out to understand the origins of the incident and the measures to be taken to prevent it from happening in the future.

Facebook and Instagram

While not cloud services per se, Facebook and Instagram rely heavily on the cloud. Earlier in 2019, a server configuration change caused a breakdown of these social networks.

For almost 14 hours, users met significant difficulties in connecting among other problems. More recently, users have complained again that Facebook is not working or that certain features such as publishing or messaging are inaccessible.

These breakdowns that have occurred throughout the year on various cloud services known around the world demonstrate that it is still unwise to entrust your data and applications to a cloud provider to this day. The zero risk does not exist, and the data may be deleted despite the promises of the suppliers. For this reason, many companies are opting for hybrid cloud and continue to store their most sensitive documents on their own servers …