1 Comment

Summary:

Amazon EC2 services running out of Northern Virginia are showing signs of wear and tear on Friday the Thirteenth.

AWS: Reinvent
photo: Barb Darrow

Friday the 13th is not starting out well for Amazon’s big U.S. East region where Amazon’s EC2 service  is showing signs of stress. Network connectivity issues there appear to be impacting Heroku, parts of Github, CMSWire and some other sites. The networking issue also impacted Amazon’s RedShift data warehousing service, Simple Messaging Services, Relational Data Service, and CloudHSM service, according to the status page.

sept13awsoutage

US East, a huge server farm in northern virginia remains Amazon’s largest and oldest facility, although Amazon spokespeople have said the infrastructure has been renovated pretty much from the ground up in the past year or so.

The Amazon status page attributes the issue to network connectivity issues in a single Availability Zone within US. East.

Things have been pretty quiet on the AWS front of late save for a minor glitch last month. Last year was a different story where several major outages including one on Christmas Eve that took down Netflix made headlines.

Stay tuned for updates.

sept13heroku

Update: According to the AWS status page the US-East-1 network issues were resolved 9:41 PDT:

10:34 AM PDT Between 7:54 AM and 9:04 AM PDT a small number of RDS instances in the US-EAST-1 Region experienced network connectivity issues in a single Availability Zone. Network connectivity has been restored and the service is operating normally.

  1. Whatever you do there (multi-zone, load-balancing, full redundancy, …) it always fails :(
    We use 3 zones, and Amazon load balancing, and for some bad reason the load-balancer still sends query to the dead zone!

    Share

Comments have been disabled for this post