So what is ClusterRunner?
ClusterRunner is a tool designed to serve a single, critical purpose - to execute tests in the fastest way possible.
What we do
- Simple, expressive interfaces to define and trigger jobs
- Manage the status of a worker-node fleet
- Real-time scaling of job-clusters based on fleet availability
- Uses historic test-execution data to optimize cluster-groupings
- Aggregate and return test result artifacts
What we don’t do
We’ve specifically avoided trying to replace functionality provided by the common CI platforms (Jenkins/Travis/etc). This includes:
- Process and present the results of a job
- Manage relationships between test jobs
- Schedule the execution of jobs
- … many other CI things.
Platforms we support
- Windows (beta)
In summary, whenever you run a test by calling a shell command, ClusterRunner can likely execute those same tests faster, safer, and more efficiently.