http://www.youtube.com/watch?v=2rKEveL55TY
talk-micro-service-architecture-fred-george#deploy-more-fewer-tests1 At 55 minutes: "We prefer the system to fail than try to do (xxx) prevention. If you think about it, tests themselves are not lean in any way whatsoever. Because tests are there just in case something goes wrong. Any time I use the word "just in case" or "it may be necessary" I've now moved outside the lean/agile world. I'm doing something almost by rote. I really want to be just in time in these sorts of things. I'd rather have a system fail fast than do (xxx) prevention. And so that's why my unit tests have gone away, because I'm deploying twice a day. I have my business metrics to tell me when I've screwed it up. I don't need to write those tests." talk-micro-service-architecture-fred-george#deploy-more-fewer-tests1
talk-micro-service-architecture-fred-george#publish-interesting-things "If you think something is interesting, then publish it." talk-micro-service-architecture-fred-george#publish-interesting-things
Would this be like logging things that are interesting so you can see interesting changes later?
Kafka - a logging thing that's a distributed hash table. (Looking back on this in 2016, I've certainly heard a lot about Kafka by now. Interesting that he was talking about it in 2012.)
http://oredev.org/2013/wed-fri-conference/implementing-micro-service-architectures