what percentage failure rate is acceptable ?

By "percentage failure rate", I mean downtime and other assorted failures in an IT operation. To cut to the chase, IT folks, what should we expect from a well run IT operation such as "Linden Lab" ? This is a sincere question seeking information.

Any takers ? I'm not in IT so I have no clue. Thanks for your time.

Steve

I would say once every Wednesday for routine maintenance and bot bootage should keep the cobwebs out.

Difficult to say, but I'll wager that the colocation/hosting company LL uses for its servers have a 99.98% uptime SLA with penalties for non-compliance.

We have a SLA (service level agreement) between the business and the IT. Lots of details such as responce times for verious severities and what exactly can cause severity levels to increase.

But the short answer you are looking of is 95% of the time we will be with in our SLA agreements -- which is yet another point in the SLA.

I would consider a 5 nines up time as the minimum for all mission critical systems. This is 99.999% free from unscheduled downtime. This can be achieved thru redundancy and careful testing before any changes are made in any hardware or software system. For those who don't think this is possible just look at finacial systems and the telco system.

Before you start throwing some of these numbers around, be aware that each company and it's IT department have done independant negotiations on the level of support the company wants to pay for. The drop from 98% to 95% was accompanied by headcount reduction as the beancounts would say.

The old saying goes "you get what you pay for" but it can be better said "you pay for what you get."

I'd say that when item delivery and Linden transaction failure rates are so commonplace that some residents actually use it to their advantage to get free stuff from merchants who are handling so many refunds and item re-sends that they don't have time to fully investigate each case, it's pretty damned unacceptable.

Considering that my monthly bill from LL is 50% higher than the mortgage on my RL house, I'd tend to agree that I'm not getting what I'm paying for in terms of reliability.

And yet I'm STILL HERE.

"nothing works!"
"grid is up!"
"but nothing works!"
"grid is still up!"
"N O T H I N G W O R K S !"
"grid is up!"

scratch that "well run"...

The asset problems over the last week are _well outside_ what I'd consider acceptable for something like SL.

Aside from that, in general, I don't think SL is too bad. They still have a lot of work to do in some areas but considering what the grid has gone thru in the 18 months, I'm probably more optomistic than most.

With something like SL, it's hard to state uptime without lots of disclaimers. Maybe you're having a horrible time on your sim but the rest of the grid is fine. Or maybe they're having (continued) asset problems but you're just hanging out chatting and don't even notice. Or maybe some router somewhere lost it's mind and just people coming from the UK and trying to get to sims hosted in Texas are having problems.. Etc, etc, etc.

Fully take your point Yosef, but what I was referring to is the SLA that LL would likely be getting from its service providers. Which will be in stark contrast to the utter bobbins we are having to endure right now.

Percentage failure rate of your tier payments allowed? 0%.

Excellent point.

Think of it this way
What % down time would you say is ok for your bank to have? Or for your CC company?
and why would that time be different for say a game?

Oh btw, blog:


Y'all were saying? *sigh*
.

LOL, did I jinx things or what?

Shame on you Wilde. hakes Fist:

You know, they could save themselves some trouble just by setting up a chron job to post a blog entry like that every day about this time...

I can't remember the last time I've experienced "down time" from a bank or credit card company. And that covers quite a few decades.

I don't recall ever using a credit card, at any time of day, and the transaction failing because the credit card company had downtime.

I've had banks that remodeled. But I didn't lose my money, or was prevented form doing transactions, during the time of remodeling. Usually I just had to go to the brank a couple of miles away.

Banks do have hours of service. But they are clearly posted hours of service. I don't know of a bank that advertises having a lobby open 24/7.

That's the thing. No company that has to actually depend on customer satisfaction for business would tolerate the incredible unexpected downtime SL has. SL falls far below any standards you'd find for any business.

Maybe if LL can only operate SL 8 hours a day, they should just open it 8 hours a day, and post hours of operation.

Availability is always as high--and only ever as high--as necessary to maximize profit for the enterprise. In industries for which downtime incurs very large costs--often where lives are at stake or similar risk of litigation--it's common to find "five nines": 99.999% availability. This applies, for example, to medical products, avionics, and telecom (E911, especially). Also, if competition on the basis of availability is a factor, the risk of lost accounts due to downtime may drive high availability targets. And if service is billed by usage--e.g., every second of downtime is revenue directly lost--that tends to push targets higher than if accounts are billed a flat rate.

But high availability is very expensive. Even though "five nines" cuts downtime only ten-fold compared to "four nines", in a complex networked system it can easily cost 100 times more to achieve. So companies don't invest more in availability than their market requires.

Translated: As long as SL residents continue to be willing to pay LL for crap, LL will continue to provide crap for sale.

They tried, but it quit working after the first day. They are looking into the situation and will get back to us at a later time.

maybe they need to use some servers you can leave and forget about pretty much.
like sql server and windows server. can't remember the last time had to do actual maintenance on one and not have it stay up to date by itself and keep fresh indexes and stuff...
stuff just works...

From midnight to 1:00 AM daily my online banking is down. The drive up bank is down for much longer periods of time

I think people are aiming to high ..

I think a nice low target should be set ..

It would be nice if SL could go 6 days out of 7 without problems.