On Fri, Aug 12, 2011 at 05:39:40AM -0400, Luke S. Crawford wrote: > On Fri, Aug 12, 2011 at 10:00:15AM +0200, Aleksandar Ivanisevic wrote: > > On 08/10/2011 06:11 PM, Brian Mathis wrote: > > >Like any agreement, it's mostly about politics and risk. There's no > > >amount of pointing to a contract that's going to prevent the client > > >from being upset when your service is down, even if you meet your > > >numbers on a yearly basis. It's during the downtime that the client > > >will form judgements about you as a provider. For that, see: > > > > > > Velocity 2010: How to turn a disaster into an opportunity - > > >http://goo.gl/BcU9h > > > > This is really a great talk, I've really enjoyed watching it and even > > started to daydream how good my world would be if I could do this, but > > unfortunately this only works when you have large public facing service > > available to millions of users that are using it around the clock. > > > > Despite all the buzz around web 2.0 and millions of users and "web > > scale", most of the services are not like that and if there is a > > slightest chance that you might fix the problem before the customer > > notices or you fix it fast enough that you can blame it on the > > "network", and usually there is, you will never get the permission to > > open your monitoring system to the customers, that's what Google and co > > are basically doing. > > Meh. I'm still under 2000 customers, and I find that being open generally > works best. I mean, sure, if you fix it fast enough that they only lose a > few packets, nobody cares. but for outages they notice? the #1 > complaint I get is that I didn't notify people soon enough. > > People really, really like being informed of what's going on, even if > you are just saying something like "I don't know what is going on > yet, but I see that there is a problem and I'm working on it" > > This is something that I simply can not convince my employee. "But > the auto response says we're working on it!" no, it doesn't. or > "but we don't have any useful information yet!" No, simply telling > the customer that we are aware of the problem and not sleeping or > whatever is useful information. > > The funny thing is that as far as I can tell, most customers prefer > to be notified of what you are doing /even if this takes time away > from actually fixing the problem/ > > So yeah. Right now, in my system, it takes a good half hour to go from > a list of down servers to a list of email addresses of users (ugh, yes. > must automate this) and I don't think they'd like me taking that much > time away from problem solving, so I usually just throw up a twitter/blog, > but yeah; not communicating enough during downtime is /by far/ the > most common complaint I get.
As one of his customers, I have been most impressed with the prgmr.com crew and their notification of various issues. Even when it hasn't directly affected me, it's kind of reassuring to see that when there is a problem, they don't try to hide it. Another company that seems to be doing pretty well in this regard is hover.com. Both prgmr and hover use twitter as a way of notifying of outages. ( And yes, I chuckled when I saw someone comment on the trend of using twitter to notify about service issues for other products, given their history ) -- Matt Okeson-Harlow http://technomage.net
signature.asc
Description: Digital signature
_______________________________________________ Discuss mailing list Discuss@lists.lopsa.org https://lists.lopsa.org/cgi-bin/mailman/listinfo/discuss This list provided by the League of Professional System Administrators http://lopsa.org/
