Watch the talk
-
-
Okay
1 reply 0 retweets 1 like -
Replying to @stefanpenner @Conaw and
@Conaw this talk resonates with me deeply1 reply 0 retweets 0 likes -
Replying to @stefanpenner @Conaw and
5 9s though, may we’ll be at odds with “ideal complexity” vs “actual complexity” trade off described in the talk…. Would love to chat about it sometime if you free
2 replies 0 retweets 0 likes -
But, nobody in this conversation is making up the phrase "five nines". That used to be a standard guarantee of reliability in the tech industry, that products achieved, and that everyone knew what it meant, in the 90s and 2000s. The fact that this has been forgotten is shocking.
1 reply 1 retweet 5 likes -
Replying to @Jonathan_Blow @stefanpenner and
(Four nines was probably more common, but the web stuff we have today is very far from four nines. 5 9s is about 2 minutes a year total downtime, and 4 9s is about 20 minutes.)
2 replies 0 retweets 2 likes -
5 9s is just shy of 6min, and 4 is around 50min. I completely agree More teams should be using 9s as a way of budgeting and planning. And knowingly choose a appropriate set of trade offs. Rather then today’s
#yolo approach.2 replies 0 retweets 0 likes -
Replying to @stefanpenner @Jonathan_Blow and
But very few services can justify the costs associated with 5 or even 4 9s, especially if they are state full and need to scale… At several places I’ve worked, we have used the above to calculate tolerable losses, and set appropriate goals and budgets to attain.
1 reply 0 retweets 0 likes -
Replying to @stefanpenner @Jonathan_Blow and
The costs are exponential, and appear not only as immediate $$ but a level of ongoing scrutiny and friction that can impede a products velocity. A vanity 9s goal, selected improperly for an evolving product that can’t justify that goal, may very well be fatal for that product.
2 replies 0 retweets 0 likes -
Replying to @stefanpenner @Jonathan_Blow and
I do hope, some serious change in today’s “foundation” can help elevate development further to help attain a better baseline, but as described in the talk, each level of abstraction brings along its own fragility. How to get the best foundation without the fragility is an open q
2 replies 0 retweets 0 likes
It is possible to do what we do today with substantially fewer levels of abstraction, and where each level is constructed to a much higher standard of quality.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.