StatusGator has been monitoring the world’s cloud services for more than 10 years now. We’ve seen outages, big and small, affect companies of all sizes for more than a decade. Yet as we close out 2025, it feels like the last 12 months brought us some of the biggest outages in the history of the internet.
In fact, by our data, this is true! Never before in history have so many huge outages taken down so much of the internet, in such a short time. Especially in the fourth quarter but throughout all of 2025, outages took down most of the major cloud hyperscalers like Google Cloud, Microsoft Azure, Amazon Web Services, and Cloudflare.
StatusGator had a front row seat to this chaos (and was even affected by some of it) because we have millions of people reporting outages to us and thousands of customers that rely on StatusGator to alert them about outages. As major cloud providers have gobbled up more and more enterprise market share, the entire world has become dependent on just a handful of providers.
Our Early Warning Signals feature exploded in importance this year, with major providers frequently delayed 10 to 60 minutes before acknowledging incidents. Early alerting of outages has become a must-have feature for IT teams, allowing them to stay ahead of disruptions and reduce support tickets with proactive communication.
Below is a timeline of the most significant outages of 2025 using data from StatusGator Early Warning Signals, outages reported to us, outages officially acknowledged by providers, and our own analytics. What remains to be seen is what 2026 will bring us: more consolidation into the large providers? Better stability overall? Or a backlash against this concentration and a move back to the principal of a distributed, open internet? Time will tell, but first let’s look back at what 2025 brought us:
Timeline of major cloud outages in 2025
OpenAI – January 23
Early Warning Signals detected the outage 5 minutes earlier.
ChatGPT exploded in popularity in 2024 and become a business-critical service in 2025. January brought the first of a seemingly endless series of disruptions to OpenAI throught the year. This outage caused widespread API errors including 503 and Bad Gateway errors. Users were unable to access the web interface or make API calls. StatusGator detected elevated user reports before OpenAI acknowledged the issue.
Box – February 19
Early Warning Signals detected the outage 25 minutes earlier.
With Box now an enterprise-level service trusted by thousands of financial firms, law firms, and many more, it’s a critical tool used by millions. In February, a maintenance operation accidentally deleted authentication keys, locking users out of Box entirely. Nearly 100,000 organizations were affected before the incident was acknowledged.
Read our blog about the Box February 19 outage
Square – February 26
Early Warning Signals detected the outage 5 minutes earlier.
Square is much more than a website hosting platform now, it’s a payment terminal used a merchants, cafes, and retailers worldwide. A certificate validation failure disrupted Square’s payment processing pipeline, forcing sellers into offline mode during peak business hours.
Check out our postmortem of the Square outage
Zoom – April 16
A DNS registry issue disrupted Zoom’s web portal and meeting join flow. Existing meetings remained active while new participants encountered internal server errors which delayed detection and acknowledgment.
This infamous Zoom outage was a stark reminder of why we always recommend hosting your status page on a different domain name: This Zoom outage took down their status page, too. Customers using StatusGator knew about the outage because alerted about it despite their status page issue.
SentinelOne – May 29
Early Warning Signals detected the outage 52 minutes earlier.
A critical infrastructure failure removed network routes and DNS resolver rules, cutting off access to SentinelOne management consoles for over five hours.
In fact, SentinelOne finally grew up and got their own official status page this year — something they never had before 2025, astonishingly enough. Because of delays in acknowledgement, we still operate our own unofficial SentinelOne status page.
Read our blog post about the SentinelOne outage
Heroku – June 10
An unintended automated OS update disrupted network connectivity for dynos across multiple regions. Heroku’s own status page was impacted, leaving customers without official updates. Heroku has since moved their status page to Salesforce’s status page yet StatusGator customers can rely on better alerting by subscribing to Heroku’s status.
Watch our video analysis of the outage
Google Cloud – June 12
Early Warning Signals detected the outage 53 minutes earlier.
A faulty API management update triggered IAM crash loops, cascading into outages at Discord, Spotify, Snapchat, and other major platforms.
One of the top cloud outages of 2025, this Google outage took down huge swaths of the internet. It should have served as an early wake up call for users dependent on a single cloud provider. Instead, it was promptly ignored and forgotten like most outages.
We wrote a blog post about the Google Cloud outage
SentinelOne – July 10
Early Warning Signals detected the outage 16 minutes earlier.
A second SentinelOne outage just on the heals of the previous. This was a even wider impact, with thousands of people reporting the issue to us. A widespread silent outage caused gateway timeouts across North America and Europe. Despite customer impact, the incident was never officially acknowledged as downtime.
Read our blog post live analysis of the SentinelOne incident
Watch on our YouTube channel
Google Workspace – July 18
Early Warning Signals detected the outage 16 minutes earlier.
A control plane failure in the us-east1 region disrupted Gmail and Drive during peak business hours for enterprise users. It also impacted thousands of schools, a key segment for StatusGator, as we serve hundreds of custom status pages for K12.
Read our Google outage blog post
Watch the live Google Workspace outage video
Starlink – July 24
A core network software failure caused a global outage, dropping Starlink traffic to roughly 16 percent of normal levels for over two hours.
StatusGator alerted about this outage long before they acknowledged the issue. Since Starlink has no official status page, they had to update their home page with the incident.
Watch the live Starlink outage video
Shopify – August 7
Shopify was affected by an outage that affected access to storefronts, another outage that StatusGator detected before it was officially acknowledged. Hundreds of users reported 500 errors with a complete inability to access the store’s website.
Watch our Shopify live outage video
SentinelOne – August 13
Yet another SentinelOne outage. This time console instability and gateway timeouts prevented security administrators from managing endpoint fleets for nearly three hours. While SentinelOne had a status page by this time, customers using StatusGator were notified about it 22 minutes before they acknowledged it, reminding users just how important it is to have a source of early alerts.
Watch the August 13 SentinelOne outage report
Starlink – August 18
A second major outage in under a month. SpaceX later confirmed an on-orbit anomaly as the cause. This outage was also reported by StatusGator before Starlink — again without a status page — acknowledged the issue on their website. If your business depends on Starlink for mobile or remote internet access, you simply must monitor Starlink status with StatusGator.
Watch the Starlink outage video
Google Meet – September 8
The most impactful of two different Google Meet outages in 2025. This one affected hundreds of StatusGator customers, especially in K12. An overloaded cache service prevented users from joining meetings across major US metros, impacting approximately 1.8 million users.
Watch the Sept 8 Google Meet outage video
Google Meet – September 26
Millions of users were affected when Google Meet was taken down worldwide due to a flawed experimentation module that caused backend contention. For a huge chunk of users (roughly 20% across the globe) the server was returning 504 errors.
StatusGator reported this outage more than 20 minutes before Google acknowledged it.
Watch my Google Meet outage video
YouTube – October 15
While not a business-critical service for most people, YouTube is a hugely popular platform and when it goes down, a LOT of people complain. The October 15 YouTube was a big one, making headlines across the globe.
A spike in playback errors primarily affected the web interface, while mobile apps remained functional, confusing users and delaying diagnosis.
Amazon Web Services – October 20
This is the outage everyone was talking about and will probably be talking about for years to come. A DynamoDB DNS race condition in us-east-1 cascaded across 141 AWS services, disrupting over 3,500 companies that StatusGator monitors. Countless organizations from the around the world were caught off guard.
Early Warning Signals detected the outage 10 minutes earlier, providing critical visibility in the early minutes of this extreme disruption.
Read our article about AWS US-East-1 reliability
Azure – October 29
Another massive outage that came barely a week after the AWS outage. If your company thought you were safe from outages because you used Microsoft instead of Amazon — think again. Early Warning Signals detected the outage 42 minutes earlier.
A configuration change in Azure Front Door disrupted global routing, grounding flights for Alaska and Hawaiian Airlines.
Google Workspace – November 12
Early Warning Signals detected the outage 62 minutes earlier.
Google’s longest reporting delay of the year. Yet another outage with Google Workspace that took down Docs, Drive, and may more Google services, impacting companies worldwide. Although not all users were affected, millions were. Google said an SSL protocol error blocked access and it took them nearly an hour to acknowledge the issue.
Our blog post about the Google Workspace outage has more details
Watch our Google Workspace outage video
Cloudflare – November 18
The second-most newsworthy outage of the entire year and other we will be talking about for months or years ahead. Cloudflare is used by countless websites around the world and their outage took down companies large and small. A malformed configuration file in Cloudflare’s bot management system crashed core proxy infrastructure, taking down a significant portion of the internet.
Read our blog post about the Cloudflare outage
Shopify – December 1
Shopify outages tend to cause wide impact to ecommerce sites and they often delay in updating their status page. In this case, StatusGator Early Warning Signals detected the outage 10 minutes earlier.
On Cyber Monday, an authentication failure locked merchants out of dashboards during the highest-traffic shopping day of the year.
Read our Shopify Cyber Monday outage blog post
Cloudflare – December 5
No good deed goes unpunished: Cloudflare tried to prevent the React2Shell vulnerability from affecting their customers and released a patch that caused HTTP 500 errors across roughly 28 percent of Cloudflare traffic. On the heals of their last outage, this was a major facepalm for Cloudflare and caused them to declare “Code Orange” to re-architect their systems.
Read our Cloudflare 2025 outage #2 blog post
Microsoft Teams – December 19
Global messaging delays impacted Teams users worldwide during the holiday season. Early Warning Signals detected the outage even though Microsoft never acknowledged it on their status page
Read our Microsoft Teams outage blog post
Watch the live analysis of the Teams outage
What 2025 made clear
- Constant consolidation into the cloud has made us all vulnerable.
- Status pages are lagging indicators. Acknowledgement often averages 10 to 60 minutes behind the actual outage.
- Silent outages are now routine, not edge cases: Many companies never publish small incidents.
- The blast radius continues to expand through shared dependencies.
- Performance degradation is downtime from a user’s perspective.
Stop being the last to know
If your first alert comes from a provider’s status page or from your users, you are already behind. StatusGator provides an independent, real-world view of service health through Early Warning Signals that detect outages when they actually start, not when they are acknowledged.
