Synthetic Monitoring
Facebook demonstrated last night that even the biggest websites aren’t immune to performance issues now and then.
According to our performance monitoring (and by many of the 900 million Facebook users), Facebook (NASDAQ: FB) began experiencing high response times and sporadic unavailability Thursday night and continuing into Friday morning.
Checks at 5:28 p.m. PST, 5:33 p.m. PST, 7:19 p.m. PST, and 7:29 p.m. PST on Thursday and again at 6:47 a.m. PST on Friday found Facebook.com completely unavailable. Users that could access the site during these periods likely experienced unusually high response times. We recorded wait times upwards of 30 seconds and even as high as 74 seconds during multiple checks on Thursday night.
The chart below shows Facebook.com’s response times (indicated by the line graph on top) and overall performance beginning at 9 a.m. PST on Thursday and continuing until about 7 a.m. PST Friday. Performance normalized shortly thereafter with no more fatal website attempts being recorded.
There’s some debate as to what caused the outage. Hacktivist group Anonymous is known for debilitating popular websites via DDoS attacks, and early tweets from organization members raised suspicion about their potential involvement. But the group later denied responsibility and a Facebook spokesperson confirmed to Computerworld that the outage was not the result of an attack.
For more on the incident, I encourage you to check out Suzanne Choney’s article on MSNBC.com, Stacey Higginbotham’s article on GigaOm, or Sharon Gaudin’s article on Computerworld, all of which include the data we recorded.