 Hello, my name is Richard Trituck working for Sunet like Mickey Nadine who presented earlier and I am going to take you on a Little journey about continuous testing at a global scale But first about Sunet. It's the Swedish University computer network. They provide a lot of network infrastructure and cloud-based services for Sweden. They have 8,400 kilometers of fiber so they like to dig their trenches and Put in fiber cables. They serve about 750,000 users among 110 organizations But sometimes you get The opportunity to build something so we build a system for scientists and so you need to model your system It's a bit of background information on why we do what we do so you have your system models where you have all the parameters of the scientific life cycle and and Build your system where you saw your data Have your life cycle management and you actually also have to face the reality of every scientific project because the money at a certain point So you build your system in a way that the universities can life cycle handle their research data Even after the project funding has run out So with this model in mind you start building and so you start building not small You start building big you build at a global scale As Mika already said earlier, we have 54 next-cloud nodes one for every Swedish University pre-built They can just log on provision their user account and use the system and We use S3 buckets as the storage Entity for handling the research data So that gives you a bit of a context of the size of The system that we have but then the scientific reality again is you need to collect data You define your primary data sources save everything into the storage as I said We use S3 and then you have the secure storage with Windows Linux Mac iOS Apple command-length lines Anything that you need Once you have that system you start expanding so your system becomes bigger and bigger You don't want to only store your files. You want to analyze your data and you want to publish your data So you integrate all the apps Once you have that expansion you start asking yourself does it actually work all the time and this is where the testing comes into Into place you start writing your tests if you're small unit test acceptance test just to see if things actually work Once you have written your tests you start testing system systematically because that's the purpose of testing, right? So you have manual testing automated testing and on the mod testing Automated testing and you hope that your automated testing will be will cover anything So you start with the acceptance testing you have capabilities stats and versions user that's life cycle handling file handling Right test for all of those you do selenium testing for no login including two-factor authentication And so on collaboration testing takes a lot of time But we do it all the time to make sure that the collaboration works and we occasionally do some load testing and Have tested with up to 6,000 users Once you do that you test automatically so you start building your Jenkins You have your worker nodes and everything will be executed all the time so you Once you have done that you start using pipelines because you realize oh this is actually a lot I don't want to click on start test if you do on-demand testing So you have your pipeline where you trigger all your tests for all nodes all the time And you get this nice matrix view about the status of your system and in the end you test a lot So we have 45 tests we create around 23,000 test points test results every single day It's about 1,000 test points per customers per day but the goal is to extend this to have even more tests in the end and You keep the developers busy. This is from our next cloud support queue so we've write a lot of tickets we find issues we solve issues and Yeah, people are smiling at me because they might realize some of the tickets we have created and In the end you help making next clouds a better product by having this type of test automation Thank you