Documentation
¶
Overview ¶
TODO Package state will contain all (ORM) entities and services to track the execution of test plans against upstream commits and corresponding Git refs (branches, tags).
It will not store metrics, artifacts, etc. Those will be stored in Elasticsearch, and the Engine will offer APIs to query both the state (sourced from this package) and the results, metrics and outputs (sourced from Elasticsearch), in a convenient and coherent way.
The most suitable database for storing this data is probably SQLite -- the data is of relational kind, and at this stage, volume and load, it's not worth adopting a standalone engine like PostgreSQL.
TODO Use jinzhu/gorm for relational ORM, and golang-migrate/migrate for schema migrations.
If we want to go the KV store route, consider badger. Beware we'll need to replicate data in different shapes to create multitude of indices (e.g. so we can query all test runs pertaining to branch X of repo Y).
Source Files