Google Cloud Platform (GCP), which affords a variety of cloud providers for numerous functions worldwide, skilled downtime on Thursday, the corporate reported.
In a standing report, Google stated that “a number of GCP merchandise are experiencing service points” throughout a number of places worldwide. The incident started at 10:51 am (PT) on Thursday and was resolved hours later, at 6:18 pm (PT), which is 11:21 am IST and 6:48 am IST, respectively.
Platforms comparable to Spotify, Replit, Discord, Claude, Snapchat, Twitch, Cloudflare, and different providers that use the Google Cloud confronted outages.
Apart from, lots of Google’s merchandise, like Google Meet, Drive, Chat, and Gmail, that depend on GCP had been stated to be affected. These merchandise use Google Cloud’s infrastructure for providers comparable to storage, computing, networking, information processing, and extra.
Amjad Masad, CEO of Replit, stated on X, “Google Cloud is having an outage and that’s taking Replit down.”
Anthropic, the corporate that develops the Claude household of AI fashions, acknowledged on a standing web page that it has recognized a number of points with its merchandise because of the outage affecting GCP. The corporate additionally talked about that picture and file uploads by way of the API and the Claude internet app had been unavailable because of the GCP challenge.
It appears to be like like EVERYTHING is DOWN
Amazon Net Companies, Anthropic, AT&T, Field, Breezeline, Calendly, CharacterAI, Cloudflare, cursor, Dialpad, Discord, DoorDash, Dragon Ball, Embark Studios, Equifax, Etsy, FuboTV, Gmail, Google, Google Cloud, Google Drive, Google Gemini, Google… pic.twitter.com/Sw5ziwbCt8— Hackmanac (@H4ckmanac) June 12, 2025
GitHub, the code internet hosting platform that additionally gives AI-assisted programming capabilities with GitHub Copilot, stated in an incident report that underlying AI fashions like Claude and Google Gemini had been dealing with points.
“Copilot is experiencing degraded efficiency,” stated GitHub.
Nonetheless, the entire above providers had been restored to normalcy after just a few hours.
Supply: x.com/GeraDeluxer
Cloudflare Acquired Hit Dangerous
Cloudflare was one of many platforms most impacted by GCP’s outage. The corporate gives web safety providers to a few of the main web sites worldwide, with round 19.3 million web sites utilizing its providers.
“The outage lasted 2 hours and 28 minutes, and globally impacted all Cloudflare prospects utilizing the affected providers,” stated the corporate in a weblog put up. It added that the outage was on account of a failure in an underlying storage infrastructure, part of which runs on a third-party cloud supplier.
A spokesperson from Cloudflare advised CNBC that its challenge stemmed from Google’s Cloud providers as properly.
“A restricted variety of providers at Cloudflare use Google Cloud and had been impacted. We anticipate them to return again shortly. The core Cloudflare providers weren’t impacted,” stated the spokesperson. The corporate’s inventory additionally fell by 5% on Thursday.
We let down our prospects at @Cloudflare right now. Our Employees KV service failed and the downstream merchandise that depend on that service had outages of their very own. We’ll publish a full postmortem quickly.
I do know that these sorts of incidents have actual and critical influence for groups…— Dane Knecht
(@dok2001) June 12, 2025
“Whereas the proximate trigger (or set off) for this outage was a third-party vendor failure, we’re in the end chargeable for our chosen dependencies and the way we select to architect round them,” stated the corporate.
Cloudflare’s Employees KV service—a distributed, serverless key-value retailer designed to be used with Cloudflare’s Employees, the corporate’s serverless computing platform—was considerably affected.
“Employees KV noticed 90.22% of requests failing: any key-value pair not cached and that required to retrieve the worth from Employees KV’s origin storage backends resulted in failed requests,” stated the corporate.
A number of Cloudflare merchandise, comparable to Entry, Gateway, WARP shopper, and Dashboard, had been additionally impacted. These merchandise are a key a part of the general infrastructure that Cloudflare gives to its prospects.
For example, Entry is a service that customers Employees KV to retailer software and coverage configuration together with person id info. Throughout this incident, Cloudflare stated, Entry failed 100% of identity-based logins for all software sorts.
The corporate claims that no information was misplaced on account of the incident.
How can Google Cloud, AWS, and Cloudflare all be down on the similar time? These corporations handle almost 90% of all web actions and functions. pic.twitter.com/a9bpAGvfTz
— Sam Abdul (@SamAbdul_) June 12, 2025
No, AWS and Azure Did Not Go Down
A number of experiences emerged that GCP’s counterparts, Microsoft’s Azure and Amazon’s Net Companies, additionally confronted outages, based mostly on info from Down Detector, a web site that checks what web sites are at present dealing with downtime. Nonetheless, neither Microsoft nor AWS have formally confirmed this.
Gergely Orosz, writer of The Pragmatic Engineer, stated that is seemingly a misunderstanding of how DownDetector checks for outages. “It can not differentiate between ‘Azure is having points’ vs ‘some websites hosted on Azure are having points’ (that may be the case on account of, eg, CF (Cloudflare)),” stated Orosz in a put up on X.
In keeping with Tom’s Information, an AWS spokesperson has confirmed that they don’t seem to be experiencing any service disruptions and acknowledged that “Down Detector doesn’t precisely replicate AWS points.”
General, Thursday’s occasions had been a stark reminder of how counting on a single cloud service supplier can convey down a few of the hottest apps on the web.
“We construct all this highly effective tech, then that one Google Cloud outage takes all of it offline,” stated a person on X. “Feels just like the universe saying: ‘Take a break. Go outdoors. You’re not in management anyway’,” they added.
The put up When Google Cloud Paused the Web appeared first on Analytics India Journal.