Oh lordy - here we go again. Salesforce.com suffered another major service outage on Tuesday, less than two weeks since it suffered its last problems.
According to Salesforce.com's system status page, seven instances went down at various points on Tuesday, starting with NA1, NA5 and NA6 in North America, followed by the CS0, CS1, CS3 and CS12 sandbox instances.
At 2:46 am PDT, Salesforce.com said that "power problems" had been detected and fixed, but the outages seem to have persisted longer. While "standard salesforce.com reporting, contacts, updates, case entry services" were available, users were warned that they might see problems with "sporadic search and file attachment performance."
At 5.30 am PDT, the Salesforce.com AppExchange was also still down.
As of 8.00 am PDT, the firm noted: "Customers may experience intermittent login issues on the NA1, NA5, NA6, CS0, CS3, and CS1 instances. We are actively working on resolution of this issue and believe we have identified the root cause. The salesforce.com Technology Team continues to work on restoring the service on CS12 instances and the salesforce.com application configurator. The salesforce.com Technology Team is still working on restoring the search infrastructure on NA1, NA5, NA6, CS0, CS1 and CS3. Search functionality on these instances is currently unavailable. We do not have an update at this time."
The problems have been worldwide - including Canada, France, the UK and India - from the complaints and calls for assistance on the Salesforce.com Twitter stream. Salesforce has asked clients to be patient and check in with the status update page.
By Wednesday morning UK time, all instances were up and running again.
Angela Eager of research firm Techmarketview said that Salesforce.com's response to the latest problems had been well handled. "Outages are a fact of IT life whether services are provided on or off premise. It is the speed and quality of response that matters and Salesforce.com does well here," she commented. "The company has a good track record on response and resolution (service availability is its lifeblood after all) and was on the case rapidly, able to restore parts of the system within a few hours so users had something to work with.
"As befits a Cloud-only, social-oriented supplier, it was communicating well, using its own Trust site which reports on system status, and pushing frequent updates out via Twitter – one a minute at one point."
But the downtime will add weight to calls for integration with customers’ enterprise monitoring and alert systems, she added. "The need will intensify as more customers start to intermingle services from different Clouds, while also integrating them with their on premise systems," argued Eager. "There is an underdeveloped market for Cloud lifecycle, monitoring and performance management that systems, service and software providers like HP, IBM and Fujitsu have the capability to push through on, as well as software providers like CA, and IT services providers.
"Cloud services is approaching the stage where these types of support tools should start to find traction."