https://www.youtube.com/watch?v=-UKEPd2ipEk
talk-kill-microservices-before-it-is-too-late#keep-architecture-rewrite-servicesAt 8:50 - they had an overall architecture that they moved to that they said they would "never need to rearchitect again". They didn't, but they did rewrite nearly all of the individual services. talk-kill-microservices-before-it-is-too-late#keep-architecture-rewrite-services
talk-kill-microservices-before-it-is-too-late#throw-away1 At 14:40 - because it's small enough you can just throw it away and restart if you have to talk-kill-microservices-before-it-is-too-late#throw-away1
talk-kill-microservices-before-it-is-too-late#disposable-components1 At 17 minutes "How do you change your software? Well, maybe you should just plan to throw it away. ... Let's just plan to throw our stuff away, so when you write it there's a little bit less stress. You can also try out new things because there's less stress about trying out the new things. And not only can we throw away the new pieces of code, but we should, because you need to be able to change your software. ... If you need to be able to change it. If you do it all the time and you plan to do it, then it becomes easy to throw stuff away." talk-kill-microservices-before-it-is-too-late#disposable-components1
talk-kill-microservices-before-it-is-too-late#stuck-on-platform-or-language At 20:30 - "When you're on a specific language or platform and you're stuck there you can't change the team" talk-kill-microservices-before-it-is-too-late#stuck-on-platform-or-language
talk-kill-microservices-before-it-is-too-late#no-shared-code1 At 22 - No shared code (shared code is also permanent code) talk-kill-microservices-before-it-is-too-late#no-shared-code1
talk-kill-microservices-before-it-is-too-late#http-and-json1 At 23 "We just have HTTP and JSON. That's how our services interact. Even though we have hundreds of services for this TODO app. It's dumb and it's inefficient. You can read it on the wire. ... you can totally understand what's happening on the services if you can see them. ... It was more important to me that the system was easy to understand than efficient. talk-kill-microservices-before-it-is-too-late#http-and-json1
talk-kill-microservices-before-it-is-too-late#monitoring-instead-of-testing1 At 26:40, gave up on tests. Unit tests are a design smell. Tests optimize for permanence. They create coupling and stasis. The mean time into resolution was the most critical thing. They got very into measuring things at a granular level. talk-kill-microservices-before-it-is-too-late#monitoring-instead-of-testing1
At 2910 there will never be an organizational structure that works. The fact that it changes is OK.