puts its reputation on (the) line puts its reputation on (the) line

Summary:'s newly-launched live system status page puts real-time information about its service availability online for all to see.


Full credit to for transparency over outages at long last with the debut of its live system status page, which reports real-time and historic status information for each of its locations.

Any Web user — whether they're an employee, a customer, a prospective customer or even a competitor —Why put such sensitive information out there on the Web for all to see? can now visit this page to examine's current track record. They'll be able to see a continuously updated panel showing whether the system is up or down or has 'performance issues', along with historic information for each day over the past two weeks. There's also an information panel that shows how many transactions on a given day in the week (currently Monday) were API calls and how many were web page views.

Screengrab of historic system status report
Amongst the bank of green lights on the historic 'system availability' table, a single red light stands out like a beacon the service disruption that affected some North America customers one week ago today. You can now click on that icon to find out what happened and what was the root cause. Scanning left along the line, you can also see the total number of transactions and their average speed for that day.

Comparing those figures to the previous Thursday suggests that customers processed about four million fewer transactions than they would have done had the systems been functioning normally, which represents about 12% of the normal total. The average response time for those that did complete was about a third worse than normal, at 584.5 milliseconds (ie a bit more than half a second rather than the usual third of a second, excluding the time taken to transit the Internet connection between salesforce and the customer). Obviously this is an average, which means customers using the affected systems probably suffered much worse response times while others elsewhere continued to experience normal service. logoWhy would put such sensitive information out there on the Web for all to see? It's all about reputation and trust (which is why Salesforce has knowingly adopted as the subdomain for its new status page). Only one thing is worse than having your system go down, and that's not knowing what's happening and whether anyone is working on getting it up again. You only have to look at some of the comments over on to see that Salesforce has lost a lot of customer confidence through the outages of the past couple of months, and it has made matters worse by not providing prompt and clear information about the nature and scale of each outage — breaking one of the most important of the five-point code of practice for service providers I outlined back in December.

Everyone knows that systems do go down from time to time (although the less often they go down, the better for everyone concerned of course). The only way to restore confidence is to make sure that, if it does happen, customers have the best possible information about what is going on. By making its systems status page public, Salesforce has gone one better. Customers now know that the company's reputation is on the line for all to see every moment that an outage continues. There is no better way of conveying to customers its commitment to fixing any service interruptions as rapidly as possible.


Phil Wainewright

About Phil Wainewright

Since 1998, Phil Wainewright has been a thought leader in cloud computing as a blogger, analyst and consultant.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Great step towards "Wainewright's Five Points"

    It's great to see (I don't actually hate them, BTW!) making this step to increasing transparancy. It certainly is tough to put that information out there (especially if it is real, accurate data, instead of phony numbers being churned out by a script... :) ). I still think that needs to get it's act together quickly, but this is a good idea. My big problem with this, from, is that it seems like what they consider to be downtime, and what their customers consider to be downtime, are two different things. I remember an article on ZDNet a few weeks ago, where claimed something like x hours of downtime during an outtage, and their customers report 4*x hours of downtime... I know my home cable modem provider's network status page only reflects issues that they have a ticket open for (they also never close their tickets, according to the page, the entire state is down and has been down for months, if you add the open tickets), as does not show real-time data.

    Justin James
  • outages versus brownouts

    I applaud for making this avaialble but as a user of the service I find the performance (response time) issues much more of an issue than availability. BTW these brownouts occur much more frequently as well. Can measure and report on the number of times people, and how many for that matter, are disrupted in their work due to brownouts as well?