One day you're a company touting an ultra-reliable service. The next day you're apologizing for launching a denial of service attack against your own customers.
Such is life for Salesforce.com.
On Wednesday, Salesforce issued a statement saying, it "has delivered 99.999% planned uptime and availability on its service to its customers" in March. The online CRM vendor even handled a whopping 40m transaction in a single day.
"There is no finish line when it comes to system reliability and availability, and our efforts to improve performance never cease," said Marc Benioff, Salesforce.com's CEO, who was more than happy to pat himself on the back. "Our community of success demands the most reliable service that we can deliver, and our outstanding performance in March is evidence of our dedication to scaling ahead of their needs."
Customers today, however, were left wondering what kind of scaling needs Salesforce was addressing when its site crapped out, making it impossible for several customers to do their work. Clients worldwide were affected for at least an hour, and North American users saw Salesforce.com go down for about four hours. Salesforce blamed the issue on a cache server overload, which it tried to fix by restarting the cache servers. The plan didn't work and forced administrators to yank the cache servers altogether. That, in turn, led to a lingering slow down in service.
Rather than touting its server virility, Salesforce was left begging for forgiveness.
"We apologize for any inconvenience these issues may have caused," the company said on a new web site meant to alert customers to service woes.
Salesforce.com set up the web site last December after an embarrassing database meltdown left customers enraged.
You can bet Salesforce.com will think twice in the future about issuing such lofty praise for itself. Best to let the customers handle that kind of thing. ®