A notable outage occurred on March 17 when Cloudflare suffered an interruption that impacted its customer in the northwest Pacific region of the US and Canada. The 33-minute outage over a one-hour period, was first observed around 10:20 a.m. EDT and appeared to center on Cloudflare nodes located in Kansas City, MO. This first portion of the outage lasted around three minutes. Fifteen minutes later there was a 22-minute incident centered on Cloudflare nodes located in Seattle, WA. Forty minutes after the outage was first observed, two more were observed, again centering on Cloudflare nodes in Seattle, WA. It was cleared around 11:25 a.m. EDT. Click here for an interactive view of the outage.
Update March 14
Outages in all three categories worldwide during the previous week were down from 385 to 281, a 27% decrease. In the US, they dropped from 168 to 137, an 18% change.
Globally the number of ISP outages decreased from 281 to 203, down 28%, and from 132 to 108 in the US, down 18%.
Cloud-provider outages fell from 26 to 11 worldwide, a decrease of 58%. In the US, they rose from four to six.
Collaboration-app network outages worldwide fell from five to two, and in the US from five to one.
On March 10, Dynamic Network Services experienced an interruption that resulted in DNS-resolution degradation on their Dyn Managed DNS service. The disruption affected users in countries including the UK, South Africa, Singapore, Australia, Ireland, France, Spain, and Portugal. The 55-minute outage was first observed around 6:40 p.m. EST and appeared to be centered on Dyn nodes located in London, UK. Twenty minutes later, a second Dyn node in Manchester, NH, showed outage conditions. The appearance of this second Dyn node coincided with a Dyn notification that their engineers had identified the issue and had implemented a fix. Twenty-five minutes into the outage, only the Dyn node located in London, UK, was exhibiting outage conditions and the number of affected services began to reduce, indicating the service was recovering. The outage was cleared around 7:35 p.m. EST. Click here for an interactive view of the outage.
On March 11, NTT America experienced an outage affecting some of its customers and downstream partners across countries including, the US, Australia, Canada, France, India, Germany, UK, Switzerland, and the Netherlands. The 20-minute outage was first observed around 3:05 p.m. EST and appeared to be centered on NTT America nodes located in Ashburn, VA, and Los Angeles, CA. Five minutes into the outage, the countries affected were reduced to just the US, UK, the Netherlands, and Germany, accessing downstream NTT networks. Five minutes later the outage cleared at the Ashburn, VA, node leaving just the node in Los Angeles, CA, exhibiting outage conditions. That outage was cleared around 3:30 p.m. EST. Click here for an interactive view of the outage.
Update March 7
Outages in all three categories worldwide fell from 393 to 385, a 2% decrease compared to the week before. In the US, they decreased from 184 to 168, 9% fewer.
Globally, the number of ISP outages decreased from 311 to 281, a 10% decline,while in the US they decreased from 166 to 132, down 20%.
Cloud provider outages worldwide increased from 22 to 26, an 18% increase, and the change in the US was an increase from two to four.
The number of collaboration-app network outages jumped from two to five, all of them in the US.
There were two notable outages during the week. On March 3, UUNETVerizon experienced an outage that impacted many of its peers and customers, including, Bank of America, JP Morgan Chase, Nomura, Samsung, and Zoom. The outage, lasting around 36 minutes over a 75 minute period, was first observed around 9:00 a.m. EST and appeared to center on UUNETVerizon nodes in Philadelphia, PA, and Ashburn, VA. This initial part of the outage lasted around 4 minutes and appeared to have a cascading impact on Cogent infrastructure located in New York, NY, and affected Cogent’s path to the JP Morgan Chase network. Approximately five minutes after the initial outage cleared, a second was observed that lasted around 13 minutes. It was observed on UUNETVerizon nodes located in Seattle, WA and Dallas, TX, as well as appearing to have a cascading impact in Level 3 Communications infrastructure located in Seattle, WA, and affecting Level 3 customers and partners in Canada. Five minutes into this second period, the Level 3 infrastructure direct outage cleared and after another five minutes, the only UUNETVerizon nodes exhibiting the issue were located in Dallas, TX. Around 9:50 a.m. EST, the third occurrence of the outage was observed 20 minutes after the second. This outage lasted around 19 minutes and was initially focused on UUNETVerizon infrastructure in Dallas, TX. Five minutes into the third period of the outage, UUNETVerizon infrastructure exhibiting problems expanded to include Seattle, WA. Approximately 10 minutes into this third period of the outage, UUNETVerizon infrastructure located in San Jose, CA was added to those in Seattle, WA, and Dallas, TX. At around 10:10 AM ET, the UUNETVerizon infrastructure located in San Jose, CA, was the only infrastructure exhibiting issues. The outage was cleared around 10:15 AM ET. Click here for an interactive view of the outage.
On March 3, PCCW experienced an outage affecting some of its U.S. customers and networks, including Flagstar Bank, Target, Bloomberg, Morgan Stanley, and Dell. The outage lasted around 31 minutes and was divided into three periods over an hour and 20 minutes. The outage was first observed around 8:45 a.m. EST and appeared to center on PCCW infrastructure located in Ashburn VA. The first period of the outage lasted around 9 minutes, before recurring 15 minutes later, again centered on PCCW infrastructure located in Ashburn, VA. This second outage lasted around 19 minutes. The third period was observed 30 minutes after the second ended and lasted around 9 minutes. The outage was cleared around 10:05 a.m. EST.
Click here for an interactive view of the outage.
Update March 1
Global outages across all three categories jumped from 279 to 393, a 41% increase over the week before. In the US outages went from 138 to 184, up 33%.
ISP outages rose from 233 to 311, a 33% increase worldwide, and from 123 to 166 in the US, a 35% increase.
Cloud-provider network outages globally jumped from 5 to 22, a 340% increase. The US accounted for two of them, up from one the week before.
Collaboration-app network outages dropped from four to two globally and from two to zero in the US.
There were three notable outages this week.
On Feb. 23, LinkedIn experienced a service disruption affecting its mobile and desktop global user base. The outage was first observed around 1:50 p.m. EST, with users attempting to connect to LinkedIn receiving server-unavailable error messages. Around 45 minutes later, services to some regions began to return, although others were still unable to use the services. After another 45 minutes, the server unavailable messages were replaced with content not available errors. The total disruption lasted around two hours, during which no network issues were observed connecting to LinkedIn web servers, further indicating the issue was application related. Service was restored around 3:40 p.m. EST. Click here for an interactive view of the outage.
On Feb. 25, Hurricane Electric experienced an outage that affected users across inthe US, UK, Australia, South Africa, New Zealand, Germany, Canada, Japan, Spain and Brazil. The outage, lasting around 36 minutes over a 45-minute period, was divided into two events. The first was observed around 1:40 a.m. EST on Hurricane Electric infrastructure in Singapore and Hong Kong. It lasted 24 minutes and initially cleared around 2:05 a.m. EST, but five minutes later, around 2:10 a.m. EST, a node in Marseille, France, was affected and created issues for around 12 minutes. The outage affected access to customer networks including Credit Suisse, Proctor and Gamble, DBS, Shell, and Bank of America. The issue was cleared around 2:25 a.m. EST. Click here for an interactive view of the outage.
On Feb. 24, Comcast Communications experienced an outage that affected peers and customers in the US, Canada, and the Netherlands. The outage, lasting around 14 minutes, was first observed around 11 p.m. EST and appeared to be centered on Comcast nodes in Newark, NJ, and affected access to customers including CBS, NBC, Bloomberg, and JP Morgan Chase. Ten minutes into the outage, the radius of the disruption expanded to include Comcast nodes in New York, NY, Dallas, TX, Chicago, IL, and Boston, MA. It affected customers in more countries including the US, UK, Australia, Canada, France, and the Netherlands. The outage was cleared around 11:15 p.m. EST. Click here for an interactive view of the outage.
Update Feb. 22
Outages overall were up 35%, from 206 to 279, compared to the week before. In the US they were up 53%, from 90 to 138.
Globally, the number of ISP outages jumped from 154 to 233, a 51% increase, and in the US, they increased from 78 to 123, up 58%.
Cloud-provider network outages dropped from 24 to five, a 79% decrease worldwide. In the US, they dropped from five to one.
Collaboration-app network outages doubled from two to four globally and from one to two in the US.
On Feb. 16, Level 3 Communications experienced a notable outage that affected multiple downstream partners and customers in countries including the US, Canada, Argentina, Mexico, and the UK. First observed around 11:40 a.m. EST, the outage lasted around 36 minutes over a one-hour period and affected access to customers including Bank of America, TiVo, and Lending Tree. The first minutes of the outage appeared to center on Level 3 nodes in San Francisco, CA, affecting only the U.S. and Canada. Ten minutes into the outage, nodes in Boston, MA, became involved, and at this point the impact spread to include the other countries. Five minutes after that, the outage in the San Francisco node cleared leaving just the Boston infrastructure in an outage condition. Twenty-four minutes after the outage was first observed it appeared to clear. Fifteen minutes later the outage reappeared, this time centered on nodes in San Francisco, Salt Lake City, UT, and Portland, OR. This second episode lasted for four minutes and was followed by two more four-minute outages each five minutes apart. The outage was cleared around 12:45 p.m. EST. Click here for an interactive view of the outage.
On Feb.18, GTT Communications experienced an outage that affected some of its partners and customers in the US. It lasted around 14 minutes, was first observed around 4:30 a.m. EST, and appeared to center on GTT nodes in Los Angeles, CA, affecting customer networks including Ford Motor Company, Guaranteed Rate and Loanet. The outage was cleared around 4:45 a.m. EST. Click here for an interactive view of the outage.
Updated Feb. 15
Worldwide outages across all three categories decreased from 267 to 206, a 23% drop from the week before. They dipped 4% in the US, from 94 to 90.
ISP outages decreased 20%, from 192 to 154, worldwide, and they stayed the same in the US at 78.
Globally cloud-provider network outages increased from 21 to 24, up 14%, and in the US bumped up from three to five.
Global collaboration app network outages remained at two for the second week in a row, while US outages dropped from two to one.
Around 4:30 a.m. EST on Feb. 11, AT&T suffered an outage centered in Washington, D.C., followed by issues in Tulsa, OK, and San Antonio, TX. that affected customers including some in the US, Germany, and Australia. Five minutes later only Washington, DC, and Tulsa, OK, nodes were involved and five minutes after that, just those in Washington, DC. Customer networks affected included J.P. Morgan Chase, Jeffries Group, Travelers Property Casualty, and ConocoPhillips. The outage lasted 14 minutes and was cleared at around 4:45 a.m. EST. Click here for an interactive view of the outage.
On Feb. 11, Cogent Communications experienced a series of outages over a period of 3 hours and 38 minutes that affected downstream providers as well as Cogent customers globally. The outage lasted a total of 43 minutes and was first observed around 1:07 a.m. EST centered on Cogent nodes in San Francisco, CA. This initial outage lasted around two minutes, and the Cogent environment was then stable for 43 minutes before experiencing a series of four-minute outages observed on Cogent nodes in Dallas, TX and Phoenix, AZ. An hour and a half after the initial outage was observed, a four-minute outage was observed centering on Cogent nodes located in Newark, NJ. The outage reappeared 25 minutes later, extending the list of affected nodes to locations including San Francisco, CA, Atlanta, GA, Cleveland ,OH, New York, NY, and Washington D.C. Less than an hour later, a 19-minute outage affected nodes including Miami, FL, Austin, TX, El Paso, TX, Houston, TX, San Jose, CA, Los Angeles, CA, and Las Vegas, NV. Customer networks affected included Ford Motor Company, Oracle, Home Depot, and TikTok. The outage was cleared at around 4:45 a.m. EST. Click here for an interactive view of the outage.
Updated Feb. 8
Globally outages in all three categories decreased from 278 to 267, a 4% decrease. In the US, outages decreased from 119 to 94, a 21% decrease.
ISP outages worldwide decreased from 214 to 192, a 10% drop. In the US, ISP outages decreased from 102 to 78, down 24%.
Cloud-provider network outages worldwide dropped from 30 to 21, down 30%. In the US they dropped from 10 to three.
Collaboration-app network outages globally decreased from five to two and stayed the same, at two, in the US.