The reliability of services delivered by ISPs, cloud providers and conferencing services (a.k.a. unified communications-as-a-service (UCaaS)) is an indication of how well served businesses are via the internet.

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 Feb. 20

Global outages across all three categories last week increased from 301 to 339, up 13% compared to the week prior. In the US, outages increased from 73 to 76, up 4%.

Global ISP outages increased from 215 to 258, up 20%, and in the US they remained the same at 53. 

Global cloud provider network outages increased from 10 to 14, and they increased from three to eight in the US.

Global collaboration app network outages decreased from five to four, and increased from zero to two in the US.

There were two notable outages:

On February 18, Zayo Group experienced an outage affecting partners and customers in the US, Canada, China, Australia, and Malaysia. The 14-minute outage was first observed around 4 p.m. EST, and appeared centered on Zayo Group nodes located in Houston, Texas; New York, New York; and Newark, New Jersey. Five minutes after being observed, the New York and Newark nodes appeared to recover. The outage cleared around 4:15 p.m. EST. Click here for an interactive view.

On February 14, Hurricane Electric experienced an outage affecting customers and downstream partners in the US, Japan, Mexico, and Hong Kong. The outage, first observed at around 8:01 p.m. EST, lasted a total of 19 minutes and was divided into two occurrences over a 29-minute period. The first occurrence, lasting six minutes, appeared centered on Hurricane Electric nodes in Paris, France. Around 8:15 p.m. EST, five minutes after the Paris nodes appeared to clear, the second occurrence was observed. It lasted around 13 minutes and appeared centered on nodes in New York, New York. As the second occurrence progressed, the number of New York nodes exhibiting outage conditions dropped, which coincided with a reduction in the number of affected regions and customers. The outage was cleared around 7:35 a.m. EDT. Click here for an interactive view.

Updated Feb. 13

Global outages across all three categories last week decreased from 331 to 301, down 9% compared to the week prior. In the US, outages decreased from 117 to 73, down 38%.

Globally, ISP outages decreased from 231 to 215, down 7%, and in the US they dropped from 85 to 53, down 38%.

Globally, cloud-provider network outages decreased from 12 to 10, and in the US they decreased from seven to three. 

Globally, collaboration-app outages dropped from 10 to five, and in the US dropped from three to none.

Global outages across all three categories last week decreased from 331 to 301, down 9% compared to the week prior. In the US, outages decreased from 117 to 73, down 38%.

Globally, ISP outages decreased from 231 to 215, down 7%, and in the US they dropped from 85 to 53, down38%.

Two notable outages:

On February 6, Microsoft experienced an outage that affected services in North America, Europe, and Asia. The outage, first observed around 10:55 p.m. EST, appeared to affect user access to Microsoft Outlook services. Microsoft confirmed that a change to some Microsoft 365 systems had contributed to the outage, and used targeted restarts to parts of their infrastructure to restore service. The bulk of the incident lasted about an hour and 39 minutes, although access issues could be seen for about four hours after that. The incident was similar to the January 25th event in terms of global reach and duration, but it did not appear to be network related, as no significant packet loss or latency or unusual routing behavior was observed. Click here for an interactive view.

On February 12, Comcast Communications experienced an outage that affected a number of downstream partners and customers across the US. The 21-minute outage was first observed around 1:30 p.m. EST and appeared to center on Comcast nodes in Denver. The outage was cleared around 1:55 p.m. EST. Click here for an interactive view.

Updated Feb. 6

Globally, outages across all three categories last week decreased compared to the week prior, from 373 to 331, down 11%. In the US they increased from 102 to 117, up15%.

Globally, the number of ISP outages decreased from 278 to 231, down 17%, and in the US they increased from 81 to 85, up 5%. 

Globally, cloud-provider network outages increased from 10 to 12 and increased from two to seven in the US. 

Globally, collaboration-app network outages increased from four to 10 and increased from one to three in the US.

Two significant outages:

On January 31, Level 3 Communications experienced an outage affecting multiple downstream partners and customers in countries including the US, Canada, Brazil, the Philippines, China, Mexico, the UK, Japan, India, Singapore, Taiwan, and Australia. The outage lasted a total of 47 minutes divided into three occurrences. First observed around 4:25 a.m. EST, the outage initially appeared centered on nodes in San Jose, California. Five minutes later, nodes in Los Angeles and San Francisco, California; Chicago, Illinois; Denver, Colorado; Dallas, Texas; and São Paulo and Rio De Janeiro, Brazil, also exhibited outage conditions. Fourteen minutes after initially being observed, the outages appeared to clear. Around 10 minutes after that, San Jose, Los Angeles, Chicago, Dallas, Denver, and San Francisco nodes appeared to exhibit outage conditions again. The second occurrence lasted 19 minutes and appeared to clear around 5:10 a.m. EST. Twenty minutes after that, a third occurrence  was observed, this time initially appearing to center on nodes located in San Jose, Los Angeles, San Francisco, Rio de Janeiro, and Sao Paul. Five minutes into the third occurrence, the Rio de Janeiro and São Paulo nodes appeared to clear, leaving the San Jose, Chicago, Los Angeles, and Denver nodes exhibiting outage conditions. The outage appeared to clear completely around 5:45 a.m. EST. Click here for an interactive view.

On February 3, Okta experienced an issue that disrupted access to its service for a number of users globally. First observed around 1:10 p.m. EST, connection requests appeared to return HTTP 403 forbidden status codes, indicating an application issue rather than internet or network problems connecting to Okta. The incident appeared to resolve for most users approximately 30 minutes later at around 1:40 p.m. EST. Click here for an interactive view.

Updated Jan. 30

Global outages across all three categories last week increased from 245 to 373, up 52% over the week prior. In the US, they jumped from 57 to 102, up 79%.

Globally, ISP outages increased from 187 to 278, up 49%, and in the us they increased from 42 to 81, up 93%. 

Globally, cloud-provider network outages increased from six to 10 outages, while in the US decreased from three to two.

Globally, collaboration-app network outages increased from one to four, and in the US they remained the same at one.

There were two significant outages.

On January 25, Microsoft experienced a significant disruption affecting connectivity to many of its services, including Microsoft Teams, Outlook, and SharePoint. First observed around 2:05 a.m. EST, the disruption appeared to impact connectivity for users globally. Around 3:15 a.m. EST, Microsoft announced a potential issue within its network configuration. Around 4:26 a.m. EST, Microsoft announced it had rolled back the network configuration change and was monitoring the services as they recovered. The bulk of the incident lasted approximately 90 minutes, although residual connectivity issues could be seen into the following day. A more detailed analysis of the outage can be found here. Click here for an interactive view.

On January 25, Cogent Communications experienced an outage affecting downstream providers as well as Cogent customers in the US, UK, and South Africa. The outage, lasting an hour and 18 minutes, was first observed around 6:40 p.m. EST and appeared centered on Cogent nodes located in New York, New York. The last 55 minutes of the outage saw a number of the New York nodes appearing to clear and coincided with a reduction in the number of affected downstream partners, customers, and regions. The outage was cleared around 8 p.m. EST. Click here for an interactive view.

Updated Jan. 23

Global outages across all three categories last week decreased from 252 to 245, down 3% from the week prior. In the US they decreased from 68 to 57, down16%.

Globally, ISP outages decreased from 189 to 187, down less than1%, but in the US they decreased from 53 to 42, down 21%.

Globally, cloud-provider outages remained the same at six, but in the US they increased from two to three.

Globally, collaboration-app network outages dropped from 12 to one and decreased from three to one in the US.

There were two notable outages:

On January 18, Time Warner Cable experienced a disruption affecting customers and partners across the US that came in two waves over the course of an hour and 10 minutes. First observed at around 1:40 p.m. EST, the first part of the outage lasted four minutes and appeared centered on Time Warner nodes in Chicago, Illinois. Fifty-five minutes after appearing to clear, the Chicago nodes again exhibited outage conditions for nine more minutes before being cleared around 2:50 p.m. EST. Click here for an interactive view.

On January 17, Qwest Communications experienced an outage affecting downstream partners and customers across the US. The 20-minute outage was first observed around 8:25 a.m. EST and appeared centered on nodes in Atlanta, Georgia. The outage was cleared around 8:50 AM EST. Click here for an interactive view.

Updated Jan 16

Global outages across all three categories last week increased from 217 to 252, up 16%, compared to the week prior. In the US they increased from 57 to 68, up 19%.

Globally, ISP outages increased from 166 to 189, up 14%, and in the US they increased from 48 to 53, up 10%. 

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

Globally, collaboration-app network outages increased from five to 12, and in the US from zero to five.

There were three notable outages:

On January 10, Cogent Communications, experienced an outage affecting multiple downstream providers and Cogent customers in countries, including the US, Singapore, Germany, and Canada. The outage, lasting a total of 43 minutes, was first observed around 1:05 a.m. EST, initially centered on Cogent nodes in Seattle, Washington. Five minutes later, the Seattle nodes appeared to clear, and nodes in Oakland, California exhibited outage conditions. Fourteen minutes after first being observed, the Oakland nodes appeared to clear. Forty minutes after that, those nodes again exhibited outage conditions. Around 2:30 a.m. EST, about an hour and 15 minutes after appearing to clear, Seattle nodes again exhibited outage conditions. The outage was cleared around 2:35 a.m. EST. Click here for an interactive view.

On January 10, Hurricane Electric experienced an outage affecting customers and downstream partners across the US, the Netherlands, Germany, Belgium, the UK, Ireland, Australia, Canada, Sweden, France, and Brazil. The 17-minute outage was first observed around 5:20 a.m. EST, apparently centered Hurricane Electric nodes in New York, New York. Five minutes later the outage included nodes in San Jose, California. Twenty minutes after first being observed, nodes exhibiting outage conditions expanded to include nodes in Los Angeles and  San Jose, California. The outage was cleared around 5:40 AM EST. Click here for an interactive view.

On January 13, Spotify experienced an outage that prevented some users globally from streaming songs or using the service. First observed around 7:40 p.m. EST, the disruption lasted around an hour and 55 minutes, with the major period of the disruption lasting about an hour and 11 minutes. During the outage, Spotify service was contactable, with a number of requests either timing out or returning “service unreachable” or “unauthorized” messages, which is indicative of backend system issues. The outage was cleared around 9:35 p.m. EST when. Around 11:16 PM EST Spotify announced that all services had been restored. Click here for an interactive view.

Updated Jan. 9

Global outages across all three categories last week increased from 120 to 217, up 81% compared to the week prior. In the US, outages increased from 31 to 57, up 84%.

Globally, ISP outages increased, from 76 to 166, up 118%, and in the US they increased from 18 to 48, up 167%. 

Globally, cloud-provider network outages dropped from eight to four, and in the US they decreased from five to one.

Globally, collaboration-app network outages increased from four to five, and in the US they dropped from one to zero.

There were two notable outages: