A History of Google Cloud and Data Center Outages

Despite Google’s reputation for reliability, outages are nothing new. Whether caused by software updates, networking issues, or – less frequently – data center fires, outages across Google Services can cause serious disruption for users.

Here’s a timeline of significant Google outages, analyzing the incidents’ causes, impact, and Google’s response:

April 2023: Google’s Rainy Day in Paris

In April 2023, Google Cloud faced a challenging day of floods, a data center fire, and other Google Cloud Networking problems, causing service disruptions across several service regions. Read more

August 2022: Google Data Center Fire

On August 8, 2022, an electrical incident caused a fire at Google’s data center campus in Council Bluffs, Iowa, injuring three employees. The fire occurred on the same day as outages of the company’s maps and search service, although Google said the two incidents were unrelated. Read more

Inside a Google data center (Image: Google)

July 2022: Google Data Centers Knocked Offline by London Heat

In July 2022, temperatures in London reached 40 degrees Celsius. Google and Oracle experienced issues with their cooling systems in the heat, causing outages and knocking some websites offline. Read more

December 2020: Google Services Including Gmail, YouTube Suffer Major Outage

Google’s outage in December 2020 affected Gmail, YouTube, Pokémon GO, Google Home, and other products globally. While outages are not uncommon, this specific outage was notable for its impact across the Alphabet portfolio. The majority of affected services were returned to functionality within an hour. Read more

Related:Data Center Disaster Recovery: Essential Measures for Business Continuity

July 2018: Google Cloud Disruption Brings Snapchat, Spotify Down

In July 2018, popular apps, including Snapchat and Spotify, temporarily went down after a failure of Google’s cloud-computing services. While it was initially unclear what had caused the outage, there were reports of an incident on the Google Cloud Status Dashboard. Read more

November 2017: Data Center Failover Error Kicks Google Cloud Services Offline

Memcache – part of Google App Engine – went down in November 2017. With Memcache being unavailable, requests went to the Datastore service, creating a surge of activity leading to errors and latency issues. Read more

August 2016: Google Explains What Went Wrong to Cause PaaS Outage

On August 11, 2016, a two-hour outage on Google App Engine affected 37% of applications hosted in its US Central region. Google blamed the incident on a traffic router software update triggering a rolling restart during standard periodic maintenance, which involved engineers shifting applications between data centers. The reduced router capacity ultimately led to overloading, and Google’s manual traffic redirection was not enough to resolve the problem until a configuration error causing a traffic imbalance was identified and fixed. Read more

Related:A History of AWS Cloud and Data Center Outages

April 2016: Google Reimburses Cloud Clients After Google Compute Engine Outage

An 18-minute outage in April 2016 affected Google Cloud Engine users across multiple regions. After the incident, Google claimed its engineering teams would be working on “a broad array of prevention, detection and mitigation systems intended to add additional defense,” and it reimbursed users up to 25% of their monthly charges. Read more

August 2015: Lightning in Belgium Disrupts Google Cloud Services

In August 2015, a series of lightning strikes in Belgium knocked some cloud storage systems offline. Reports had initially stated that lightning had struck electrical systems at one of its data centers in the small town of St Ghislain, but a spokesperson later confirmed a local utility grid had been hit. Read more

March 2015: Google Traces Cloud Outage to Faulty Patch

In March 2015, for the second time in a month, Google Compute Engine suffered an outage, with some users experiencing disruptions for up to 45 minutes. It was a partial outage, meaning some users weren’t impacted, some saw a slowdown, while others experienced issues contacting their cloud VMs. Google identified a patch problem as having caused the issues. The configuration change was tested before deployment but affected some VMs when live. Read more

Related:Incident Response: Lessons Learned from a Data Center Fire

February 2015: Google Compute Engine, AOL Suffer Early Morning Outages

February 19, 2015, saw two outages on the same day: Google Compute Engine was down across multiple regions for around an hour, and AOL experienced an extended outage that lasted most of the morning. Google blamed the Google Compute Engine incident on network issues, which caused connectivity loss across many zones. AOL’s email service was resolved after a morning of issues, with AOL slow to disclose the cause of the problem – some claimed there had been a network issue. Read more

October 2014: Several Google Cloud Services Experience Downtime

Google Cloud Services users experienced issues with Gmail, Google Hangouts, Google Analytics, and Google email security service Postini in October 2014. While the incident affected the majority of users, it was resolved relatively quickly. Read more

January 2014:  Gmail Web App Outage

On January 24, 2014, the widely used app Gmail went down due to an internal bug generating an “incorrect configuration.” Read more

December 2012: Load Balancer Misbehavior Cited in Google Outage

In December 2012, an incident report confirmed the cause of a recent Gmail outage to be a software update that caused a networking issue, specifically in Google’s load balancers. Google explained that a “bug in the software update caused it to incorrectly interpret a portion of Google data centers as being unavailable.” The outage had caused issues for users accessing Gmail, and many Chrome users also experienced browser crashes. Read more

February 2010: When the Power Goes Out at Google

After a power outage in February 2010, Google shared a series of steps it would be taking to address the incident. Google committed to additional scheduled drills for on-call staff, regular audits of operations documents, a clear policy framework for emergencies, and a major infrastructure change in App Engine. The outage had caused more than two hours of downtime for Google App Engine. Read more

September 2009: Router Ripples Cited in Gmail Outage

On September 1, 2009, a Gmail outage meant users were unable to access Gmail via the web interface. Google stated that the cause was its underestimating the load that routine maintenance on some Gmail servers would place on supporting routers. Google fixed the problem by bringing additional routers online and said it then increased Gmail’s router capacity and is taking further steps to avoid a repeat of the incident. Read more

July 2009: Google App Engine Hit By Outage

On July 2, 2009, Google App Engine experienced high latency and error rates, causing hours of performance issues – all applications accessing the Datastore were affected. Read more

May 2009: Rolling Outage for Google

On May 14, 2009, an error in one of Google’s systems caused traffic to be directed through Asia, creating a traffic jam. The incident affected about 14% of users, with issues reported on Google News, Gmail, and Google Calendar, among other services. Read more

February 2009: Gmail Outage Focused on European Network

A Gmail outage on February 24, 2009, was caused by disruptions in its European data centers. Unexpected issues with a software update resulted in over two hours of downtime for Gmail users. Read more

August 2008: Gmail Service Outage

August 11, 2008, also saw a Gmail service outage: many Gmail users were unable to access their email due to an issue in the contacts system used by Google that prevented Gmail from loading properly. Read more

June 2008: Google App Engine Outage

On June 17, 2008, Google App Engine, the utility computing platform for developers, experienced several extended outages during which a significant percentage of requests resulted in errors. The errors were related to the Datastore. Read more