The repository contain an auto scaling framework, simulator and runtime scaling accross multiple execution environments for an application that submit jobs. These should be well defined jobs that execute and complete. The framework is not designed to be used with continuously running jobs, such as web servers.
The framework was designed for the META-pipe backend and the implemented estimator and simulator / runtime endpoints support only META-pipe jobs.
To generate the database tables, use the provided "create_database.sql" script.
The application can be launched in two modes, one for the simulator and one for the runtime service. (Note that the runtime is not fully implemented). The updateDB flag can be set at runtime to initialize the database with META-pipe jobs. This should be done at least once to download the estimator training data.
- gorilla/mux https://github.com/gorilla/mux
- lib/pq https://github.com/lib/pq
- segmentio/ksuid https://github.com/segmentio/ksuid
- sajari/regression https://github.com/sajari/regression