Weekly internet health check, US and worldwide

ThousandEyes is monitoring how these providers are handling the performance challenges they face. It will provide Network World a roundup of interesting events of the week in the delivery of these services, and Network World will provide a summary here. Stop back next week for another update, and see more details here.

Updated Oct. 24

Global outages across all three categories last week increased from 283 to 374, up 32% compared to the week prior. In the US, they increased from 72 to 94, up 31%.

Globally, ISP outages jumped from 194 to 293, up 51% while in the US they increased from 55 to 72, up 31%.

Globally cloud-provider network outages jumped from six to 10, and in the US increased from one to four.

Globally collaboration-app network outages decreased from nine to seven, and in the US decreased from six to four.

Two notable outages:

On October 19, LinkedIn experienced a service disruption affecting its mobile and desktop user base. The disruption was first observed around 6:34 p.m. EDT, with users attempting to post to LinkedIn receiving error messages. The total disruption lasted around an hour and a half during which no network issues were observed connecting to LinkedIn web servers indicating the issue was application related. The service was restored around 7 p.m. EDT.

On October 22, Level 3 Communications experienced an outage affecting downstream partners and customers in the US, Canada, the Netherlands, and Spain. The outage lasted a total of 18 minutes divided into two occurrences distributed over a 30-minute period. The first occurrence was observed around 12:35 a.m. EDT and appeared centered on Level 3 nodes in Chicago, Ilinois. Five minutes later, nodes in St. Louis, Missouri, also exhibited outage conditions. Ten minutes after the outage appearing to clear, the St. Louis nodes began exhibiting outage conditions again. The outage was cleared around 1:05 a.m. EDT. Click here for an interactive view.

Updated Oct. 17

Global outages across all three categories last week decreased from 328 to 283, a 14% decrease compared to the week prior. In the US, outages dropped from 101 to 72, down 29%.

Global ISP outages decreased from 239 to 194, down19%, and in the US decrease from 76 to 55, down 28%.

Global cloud-provider network outages dropped from 12 to six, while in the US they dropped from six to one.

Global collaboration-app network outages decreased from 10 to nine, and from seven to six in the US.

Two notable outages:

On October 10, Microsoft experienced an outage affecting downstream partners and access to services running on Microsoft environments. The outage, which lasted 19 minutes, was first observed around 3:50 p.m. EDT and appeared centered on Microsoft nodes in Des Moines, Iowa. Ten minutes after that, nodes in Los Angeles, California exhibited outage conditions and appeared to clear five minutes later. The Des Moines outage was cleared around 4:10 p.m. EDT. Click here for an interactive view.

On October 12, Continental Broadband Pennsylvania experienced an outage affecting some customers and partners across the US. The outage lasted around 49 minutes in total, divided into four occurrences distributed over a period of an hour and 45 minutes. The first occurrence was observed around 11:10 p.m. EDT, lasted 23 minutes, and appeared to focus on Continental nodes in Columbus, Ohio. The first occurrence appeared to clear around 11:35 p.m. EDT. Five minutes later, Cleveland, Ohio, nodes exhibited outage conditions before clearing after four minutes. Fifteen minutes after that, the Columbus nodes once again exhibited outage conditions. The outage was cleared around 12:55 a.m. EDT. Click here for an interactive view.

Updated Oct. 10

Global outages across all three categories last week increased from 301 to 328, up 9% compared to the week prior. In the US they decreased from 107 to 101, down 6%.

Globally ISP outages increased from 233 to 239, up 3%, and in the US they decreased from 78 to 76, down 3%.

Globally cloud-provider network outages doubled from six to 12, while in the US they remained the same at six.

Globally and in the US, collaboration app network outages remained the same with 10 outages globally and seven in the US.

Two notable outages:

On October 4, Deft experienced an outage affecting some of its customers and downstream partners across the US, Brazil, Germany, Japan, Canada, India, Australia, the UK, France, and Singapore. The outage lasted around an hour and six minutes in total, divided among four occurrences over a period of an hour and 30 minutes. The first occurrence was observed around 5:25 a.m. EDT and appeared to center on Deft nodes in Chicago, Ilinois. It lasted 14 minutes and appeared to clear around 5:40 a.m. EDT. Five minutes later, a second occurrence lasting 19 minutes was observed with Chicago nodes exhibiting outage conditions. The third occurrence lasting 24 minutes was observed around 6:10 a.m. EDT, again centered on Chicago nodes. Ten minutes later they appeared to clear, but began exhibiting outage conditions again. The outage was cleared around 6:55 a.m. EDT. Click here for an interactive view.

On October 5, TATA Communications America experienced an outage affecting downstream partners and customers in the US, the UK, France, Turkey, the Netherlands, Portugal, India, and Israel. The outage, lasting 9 minutes in total, was first observed around 9:25 a.m. EDT and appeared initially to center on TATA nodes in Newark, New Jersey, and London, England. Five minutes into the outage, the Newark and London node outages were joined by nodes in Marseille, France. The outage was cleared around 9:35 a.m. EDT. Click here for an interactive view.

Updated Oct. 3

Global outages across all three categories last week decreased from 304 to 301 compared to the week prior, while in the U.S., they increased from 90 to 107, up 19%.

Globally, ISP outages increased from 232 to 233, while in the US, they increased from 65 to 78, up 20%.

Globally, cloud-provider network outages dropped from 14 to six and remained the same in the US at six.

Globally, collaboration-app network outages jumped from three to 10 outages, while those in the US rose from three to seven.

Two notable outages:

On September 29, Microsoft experienced an outage affecting some downstream partners and access to services running in Microsoft environments. The 33-minute outage was first observed around 7:10 a.m. EDT and appeared centered on Microsoft nodes in Washington, DC. Around 7:15 a.m. EDT nodes in Ashburn, Virginia also exhibited outage conditions. A further 15 minutes later, New York, New York, nodes also exhibited outage conditions. The outage was cleared around 7:45 a.m. EDT. Click here for an interactive view.

On October 1, Cogent Communications experienced a series of outages over a period of 35 minutes affecting downstream providers in the US, France, Singapore, Germany, the UK, Canada, and Mexico. The outage, lasting a total of 17 minutes, was first observed around 3:20 a.m. EDT centered on Cogent nodes in Oakland, California, and Washington, DC. Around 3:25 a.m. EDT, the Washington, DC, nodes appeared to clear, but nodes in Salt Lake City, Utah, showed outage conditions. This lasted around nine minutes, and the Cogent environment was then stable for 15 minutes before experiencing an eight-minute outage with Oakland and Los Angeles, California, nodes exhibiting outage conditions. Five minutes into the second occurrence, the nodes in Washington, DC, New York, New York, Houston, Texas, San Francisco, California, and Bilbao, Spain, in exhibited outage conditions. The outage was cleared around 3:55 a.m. EDT. Click here for an interactive view.

Updated Sept. 26

Global outages across all three categories last week decreased from 347 to 304, down12% compared to the week prior. In the US, outages decreased from 108 to 90, down 17%.

Globally, the number of ISP outages decreased from 252 to 232, down 8%, and from 76 to 65 in the US, down 14%.

Globally, cloud-provider network outages decreased from 15 to 14, and in the US decreased from eight to six.

Globally collaboration-app network outages occurred only in the US and decreased from nine to three.

On September 23, NTT America experienced an outage affecting some customers and downstream partners across countries including the US, the Netherlands, Hong Kong, Switzerland, and Japan. The 18-minute outage was observed around 3 p.m. EDT and appeared to center on NTT nodes in New York, New York. Around 10 minutes later, nodes in Ashburn, Virginia, also began exhibiting outage conditions. The outage was cleared around 3:20 p.m. EDT. Click here for an interactive view.

On Sept. 25, TierPoint experienced an outage affecting some customers and downstream partners across the US and Canada. First observed around 1 p.m. EDT, the outage, lasting a total of 41 minutes over a 55-minute period, appeared to center on nodes located in Nashville, Tennessee. Fifteen minutes later, outage conditions appeared to exist in nodes in Raleigh and Charlotte, North Carolina, and 10 minutes later, they appeared to recover. Around 1:35 p.m. EDT, the nodes located in Nashville appeared to recover, but exhibited outage conditions five minutes later. This second occurrence lasted four minutes before appearing to clear and then once again began exhibiting outage conditions. The outage was cleared around 1:55 p.m. EDT. Click here for an interactive view.

Updated Sept. 19

Global outages across all three categories last week decreased from 414 to 347, down 16% compared to the week prior. In the US, they decreased from 133 to 108, down 19%.

Globally, ISP outages decreased from 304 to 252, down17%, and in the US increased from 73 to 76, up 4%.

Globally, cloud-provider network outages dropped from 27 to 15, and in the US from 18 to 8.

Globally, collaboration-app network outages increased from six to nine, and in the US increased from three to nine

Two notable outages:

On September 14, Level 3 Communications experienced an outage affecting downstream partners and customers across the US. The 29-minute outage was first observed around 1:35 a.m. EDT and appeared to be centered on Level 3 nodes in Seattle, Washington. The outage was cleared around 2:05 a.m. EDT. Click here for an interactive view.

On September 15, at 11 a.m. EDT, Zoom Communications experienced an issue affecting users globally for about 24 minutes. The outage appeared to affect users’ ability to start and join meetings, but network connectivity to Zoom did not appear to experience any significant issues. Around 11:22 a.m. EDT, Zoom announced it was aware of the issues and were investigating. Zoom announced the outage was fully resolved around 11:49 a.m. EDT. Click here for an interactive view.

Updated Sept. 12

Global outages across all three categories last week increased from 327 to 414, up 27% compared to the week prior. In the US, outages increased from 87 to 133, up 53%.

Globally, ISP outages increased from 260 to 304, up 17%, and in the US they increased from 64 to 73, up14%.

Globally, cloud provider network outages jumped from four to 27, up 575%, while in the US they increased from two to 18, up 800%.

Globally, collaboration app network outages decreased from eight to six outages, down 25%, and in the US decreased from five to three.

Two notable outages:

On September 6, Hurricane Electric, experienced an outage affecting customers and downstream partners across countries including the US, Spain, Hong Kong, Canada, Malaysia, and the United Arab Emirates. The 29-minute outage, first observed at around 8:50 p.m. EDT, initially appeared centered on Hurricane Electric nodes in New York, New York. About five minutes later, the New York nodes appeared to recover, but nodes in Singapore and Marseille, France, exhibited outage conditions. Around 9 p.m. EDT, New York, Singapore, Marseille, France, nodes and those in Paris, France, all exhibited outage condition, boosting the number of partners and regions impacted. The outage was cleared at around 9:20 p.m. EDT. Click here for an interactive view.

On September 7, Microsoft experienced an issue that affected connections and services leveraging its Azure Front Door (AFD) platform. First observed around 12:20 p.m. EDT, the major portion of the outage, lasting around 50 minutes, and appeared to affect users’ ability to connect and access Microsoft cloud services that use AFD. Network connectivity to AFD edge locations did not appear to experience any significant issues throughout the outage. Microsoft’s preliminary analysis reported that the disruption appeared to be the result of an unusual spike in traffic, causing multiple environments managing the traffic load-balancing to go offline. Microsoft remediated the residual impact and announced the outage fully cleared at 3:55 p.m. EDT. Click here for an interactive view.

Updated Sept. 5

Global outages across all three categories last week decreased from 363 to 327, down 10% compared to the week prior. In the US, outages decreased from 91 to 87, down 4%.

Globally, ISP outages increased from 250 to 260, up 4%, and in the US from 57 to 64, up 12%.

Globally, cloud-provider network outages dropped from 11 to four, and from six to two in the US.