The new Salesforce approach to outages

Stung by bad publicity over service outages, the on-demand CRM company hopes to show it's doing a good job.

Colin Barker Special to CNET News
3 min read
Salesforce.com has responded to some of the criticism it has suffered after service outages left customers without their CRM services for as long as an hour and a half.

These problems hit the headlines just before last Christmas when Salesforce faced its biggest outage to date. Since then, the service has suffered intermittent faults, prompting the company to launch its own service to track outages.

The service, Trust.salesforce.com, launched late last month, tracks the system performance of Salesforce's four main servers--two in the U.S., one covering Europe, the Middle East and Africa, and one covering the Asia-Pacific region. At the end of each day it publishes a performance summary, with a traffic-light color-coding system using green, amber and red to indicate good service, minor problems and outages, respectively.

"When we had a problem at Christmas, the biggest thing that people said to us was, 'Why didn't you tell us?'" Phill Robinson, chief marketing officer for Salesforce, told ZDNet UK this week.

"Well, we do what the customer wants, and now our customers can see how the site is doing and they can also see the state of each individual server," he said.

At the time of writing, the service showed that two significant outages had occurred on one service in the U.S.--one between 10:54 a.m. and 12:15 p.m. PT on Feb. 9 and one between 6:28 p.m. and 7:30 p.m. PT on Feb. 16.

The first was the most disruptive, and happened when "a primary hardware server in our cluster failed and one of our North American (NA1) servers did not automatically recover. This required a manual restart of the NA1 database, which completed at 12:15 PT," according to the Web site.

Robinson insists that the launch of the site was not a "knee-jerk" reaction to Salesforce's recent problems.

"It is a fact that our profile is such that if something happens then it is news," he says. "It's news just because it has happened to us. We don't mind that. What I would like to ask is, do our competitors do that? I am not sure they are. The fact is that systems do have problems and outages from time to time," said Robinson.

Customers in the U.K. and the rest of Europe escaped last December's outage, according to Salesforce, but it riled a voluble section of the U.S. user base, which became angry that a system they had grown to trust had let them down.

Salesforce claims an availability rate of between 99 percent and 100 percent. Yet some of the customers that complained about the December outage also said that smaller, less-disruptive outages occurred more frequently than they anticipated.

Salesforce is trying to address this issue as well with the traffic-light system within Trust.salesforce.com, whose performance indicator will go "yellow for issues that last longer than 10 minutes, and red for disruptions that last longer than 30 minutes."

The system also identifies "incidents of less than 10 minutes"--of which there were eight during February.

Colin Barker of ZDNet UK reported from London.