pipeline-metrics-collector-client

Python-based worker collects data from gitlab API about pipeline duration and e2e jobs data and send request to pipeline-metrics-collector API

To separate the time of running tests in a job from the time of setting the environment ("up" from "e2e), its trace is parsed by time (commands for running tests and raising are wrapped in time)

Total duration (time) in statistics = time elapsed from the moment the pipeline started until the the statistics were collected. That is, if one of the jobs failed and it was not immediately blocked, the total time may skyrocket 🚀

How to use

0. Run pipeline-metrics-collector server

1. Build image for client-worker

     docker image build -t metrics-collector-client .; 

2. Run worker

   For specified pipeline use parameter --pipeline-id, for multipal pipelines use --per-page and --page parameters

    docker run metrics-collector-client \
    --save-endpoint http://web:5000/save_metrics \
    --private-token <PRIVATE-TOKEN> \
    --project-id 123 \
    --project-name test/project \
    --stage-build build --jobs-build build-e2e \
    --stage-e2e test \
    --jobs-e2e-blacklist test:e2e:lint coverage lint \
    --pipeline-id 1  # for specified pipeline
    # --per-page 10 --page 1

   For local debug (server and client running on same machine) use --network host parameter for run

     docker run --network pipeline-metrics-collector_default metrics-collector-client ...