Tuesday will be remembered as the day the internet broke — before swiftly being fixed again. Early in the morning, websites including Amazon, Reddit, Spotify, Ebay, Twitch, Pinterest and, unfortunately, CNET went offline due to a major outage at a service called Fastly. Everywhere you looked, there were 503 errors and people complaining they couldn’t access key services and news outlets. It all demonstrated just how much of the internet relies on this largely unheard-of cloud computing service.
In all, it looks like a huge range of Google services were down for about an hour today. That hour that crossed into business operation times in multiple markets, leading to a slight drop in pre-market trading for parent company Alphabet.
It’s also an alarming reminder of just how far Google reaches, and how many of our services — productivity, entertainment, and home/utility — are tied up with a single, proprietary provider.
The outage hit the entire nation of 21 million people at around midday, and although power returned to some parts of the country after six hours, Colombo was in darkness as night fell.
Egyptian authorities said Monday they had fined the national branch of mobile phone giant Vodafone 500,000 euros over a coverage outage last week just ahead of the Eid al-Fitr holiday.
The National Telecommunications Regulatory Authority (NTRA) said it made the „unprecedented“ move after services dropped „for several hours, in multiple regions“ on June 3 just before the holiday which marks the end of Ramadan, the Muslim holy month of fasting.
Data from the incident lets us rule out a global routing issue as the underlying cause, and NetBlocks also hasn’t observed the kind of spike in latency prior to the outage usually associated with a DDoS attack. This points to a different kind of attack or technical issue within Facebook’s data centres. Analysis of data shows partial failures starting some hours prior to the initial NetBlocks alert.
Reports from DownDetector suggested that the problem was global, and peaked between 9:00 and 15:00 PDT for all three services, before tailing off, though there are still residual reports of problems with all three applications.