Does some stuff and need to produce one or multiple txt files for matrix cases
At the end of job such files are processed and composed to proper matrix output in json format
Main matrix job. Does the work for all combinations of matrix cases. Uploads per-job artifacts with results (if needed)
Just sample of aggregation. For presented workflow will produce some kind of that:
=== BEGIN OF OUTPUT ===
7caedb7a196cbd6c5d11191d5a6ed464ea3e5c0c
Run job1 in runner
Run job3 in runner
Run job2 in runner
=== END OF OUTPUT ===