Hi HN,
I turned my open source ruby gem knapsack into indie SaaS product. I was inspired by sidekiq gem. Here is the short story:
In 2014 I had a problem with large test suite in my customer Ruby on Rails project. I developed the open source gem knapsack to split tests among CI nodes in parallel.
In 2015 I had an idea to turn it into small SaaS product and by the end of the year, I had a first early adopter Buildkite team.
This year I released a new solution to dynamically split RSpec tests among CI nodes and I tested it with a few companies like Pivotal, VTS, Namely.
On one hand would be great to live in a world with small and fast test suites but on the other hand, we know how it is in the reality. There are large projects with dozen minutes or several hours test suite time execution.
I also wrote the article inspired by work we've done with one of my customers:
http://docs.knapsackpro.com/2017/auto-balancing-7-hours-test...
I'm curious if there are similar problems with getting optimal time execution for large test suite across multiple CI nodes in projects done in other languages than Ruby. Have you experienced that problem? I'm thinking about building my tool in a different language and would be nice to test solution on the real project.
I'd love to hear your story about working with big test suites and what languages & tools you use. :)
