Salesforce suffered a partial outage in one of its European instances caused by problems in its storage layer that disrupted its services for some of its clients in the region.
Updates posted by Salesforce on its Trust community website said that the storage failure caused a break in communication to the storage tier, which had the knock-on effect of taking Salesforce’s cloud-powered customer relationship management software offline for some European users.
“The Salesforce EU2 instance(s) is currently experiencing a service disruption. During a service disruption, end users are unable to access the service,” Salesforce said at the time of the outage, which lasted around 10 hours before being fixed.
Salesforce was forced to report that its EU2 instance was suffering degraded performance even after the services were restored. Customers were able to access Salesforce services but some of the features were not working or performing correctly.
The storage layer is now back up and running as normal and Salesforce services have been restored. But the incident shows that, despite the growth and spread of cloud services, they are yet to be a wholly infallible way to deliver software, platforms and infrastructure from an outsourced location.
Salesforce maintains multiple instances in the Europe, Middle East and Africa region, so one instance going down is not a critical problem, but its affect will have been felt by customers with services linked to the EU2 instance.
The company’s cloud services handle hundreds of millions of transactions a day, and even a relatively minor service outage can cause disruption for many customers.
V3 contacted Salesforce for more information and comment on the outage, but the company has yet to respond.
Cloud outages are not an uncommon problem for technology companies, and V3 has a roundup of the top 10 outages that have blighted the likes of Twitter, iTunes and Amazon Web Services.
There is no silver bullet for such outages, but savvy cloud companies can design their infrastructure to effectively flow around hardware failures and start backup systems before services are disrupted.
Rigorous testing of backup systems and the ability to cope with outages is also a way to minimise their impact, and in turn make companies feel more comfortable with their move to the cloud.
V3 will host a Cloud & Infrastructure Live online event on 20-21 April. Register now to hear more about issues concerning data centres and the cloud.
Not all loose ends tied yet, admits Bain backer SK Hynix
It's Stack Overflow's second calculator, and first for external devs
Theresa May always the keenest cabinet voice in favour of draconian online censorship, surveillance and controls
No need to waste time on Google launch planned for 4 October