There are three components:
- Collector Moves data from yet uncharacterized substreams embedded in a single kafka stream to the Storage. This role is implemented by the ingress component.
- Storage Database which stores timeseries data until enough has been collected for processing and a list of processed timeseries. This role is implemented by the postgres component.
- Analyzer Periodically scans the Storage for timeseries that are ripe for processing, processes them and marks them as processed. This role is implemented by the process component.
Additionally, there is a kafka component for testing purposes. How these components are connected can be configured through environment variables. Defaults exists for each environment variable in the ingress and process. The first thing these components do is to print out their configuration along with the names of the environment variables relevant of changing it. Dockerfiles exists for all components.