Salesforce.com, which provides on-demand customer relationship management software, suffered another service outage April 6—one that lasted a good portion of the day, according to users who have contacted eWEEK.

The outage began shortly after 8 a.m. PDT on April 6 and continued sporadically throughout the day, leaving users unable to access their Salesforce.com applications for hours at a time.

At least one user noted at 4:12 pm PDT that the North American servers were down for the beginning part of the day, while another user told eWEEK that servers were down for more than five hours.

Resource Library:

This is not the first time Salesforce.com has had issues with service levels. The company has had several high-profile system outages, including a major outage on Dec. 20, 2005, that brought the company's servers down for six hours. Salesforce.com officials confirmed a second "minor" outage on Jan. 6, 2006, that affected its European, Middle Eastern and Africa regions, and a third outage was reported Jan. 30 that lasted about four hours, according to Salesforce.com customers.

Over a course of a dozen interviews with Salesforce.com customers and partners in mid-January, eWEEK learned that the company had had sporadic outages over the past year. Problems have been blamed on several circumstances, including the switch to a new data center, an undocumented bug that CEO Marc Benioff said caused the Dec. 20 problems, and the introduction of the Winter '06 release, according to some customers.

Users who tried to log into their Salesforce.com applications April 6 found this message: "The Salesforce.com servers are temporarily unable to respond to your request. We apologize for any inconvenience."

It wasn't until 12:49 a.m. PDT that the Salesforce.com system status page—put into place earlier this year to assure customers were updated in real time regarding any problems—showed that there was actually an issue with the North American servers.

Click here to read more how Salesforce.com's growing pains are irking customers.

"An issue in the cache server has caused the performance degradation and intermittent availability on the NA 1 service," read the error message that showed up three hours after customers first noticed problems. The warning enumerated the steps taken to address the issue: At 8:11 a.m. PDT, the cache servers became overloaded, and the operations team diagnosed the problem as a software issue, rolled back the code and restarted the cache servers. "The team determined the rolled-back code was not the remedy; Operations removed cache servers from the configuration," said the message on the site.

Finally, the team stabilized availability with the "side effect of slower page times."

At least one user who contacted eWEEK suggested that it's Salesforce.com's AppExchange integration platform that is causing the issues by "increasing the server load significantly."

Announced in September 2005, AppExchange is a real differentiator for Salesforce.com. As an applications marketplace where users can share and integrate Salesforce.com and third-party applications, AppExchange has propelled the company into a world beyond CRM.

Company officials have not confirmed that there are any issues with AppExchange.

Check out eWEEK.com's for the latest news, reviews and analysis about customer relationship management solutions.