HewlettPackard/zing-stats

Name: zing-stats

Owner: Hewlett Packard Enterprise

Description: Tool for generating summary stat reports and graphs from Gerrit (https://www.gerritcodereview.com/) and GitHub Enterprise review and pull requests data using https://plot.ly/.

Created: 2017-08-30 13:47:21.0

Updated: 2018-05-10 07:51:00.0

Pushed: 2018-05-01 13:25:40.0

Homepage:

Size: 278

Language: Python

GitHub Committers

User	Most Recent Commit	# Commits

Other Committers

User	Email	Most Recent Commit	# Commits

README

© Copyright 2017 Hewlett Packard Enterprise Development LP

zing-stats

Introduction

Tool for generating summary stat reports and graphs from Gerrit (https://www.gerritcodereview.com/) and GitHub Enterprise review and pull requests data.

Terminology:

• A change consists of 1 or more revisions (aka patches)
• Each time CI is triggered on a change, a CI run occurs
• A CI run consists of one or more CI jobs e.g. foo-copyright, foo-check and so on

For now, zing-stats expects job comments in the format printed by OpenStack Zuul. This needs to be more customisable in future. For now, if your CI system uses a different format, you'll need to modify the hardcoded regexs.

Screenshots

Example output with projects tab (–project-map specified)

Example output where CI jobs include duration info

Example output without projects (no projects tabs)

Example output of detailed stats section

Installation

Using virtualenv

clone <url to zing-stats repo>
 apt install virtualenv
ualenv ~/venv/zing-stats
venv/zing-stats/bin/activate
install -r requirements.txt

Running

Running as script

ng_stats.py --gerrit-url <gerrit url> --github-url <github enterprise url> --github-token <github token> -o <output directory>

The team names used in zing-stats reports are read from projects.json

Running from docker hub

(assumes projects.json is in your pwd, adjust the -v as neccesary if not)

er run -v $(pwd)/projects.json:/projects.json -e GERRIT_URL=<gerrit url> -e GITHUB_URL=<github enterprise url> -e GITHUB_TOKEN=<github token> zingstats/zing-stats

e.g.

er run -v $(pwd)/projects.json:/projects.json -e gerrit_host=https://review.openstack.org/ -e github_host=https://github.com/ zingstats/zing-stats

Running in docker compose

Uses docker compose to stand up dedicated containers for generation and reporting.

1. Build the environment,

   er-compose build --force-rm --no-cache
   

2. Create a docker-compose environment file (zing_stats.env) with your zing-stats arguments e.g.

   IT_URL=https://gerrit.example.net
   UB_URL=https://github.example.net
   UB_TOKEN=345sdfe
   _PROJECTS=/var/tmp/projects.json
   

3. Run the environment (detached)

   er-compose up -d
   

4. Attach to running container

   er-compose exec zing-stats /bin/bash
   er-compose exec zing-stats-web /bin/ash
   

5. zing-stats output at http://localhost:8172/last_7d/ (may take some time depending on number of configured projects)

Testing

on setup.py test

Design

General strategy

1. Pull changes data from Gerrit (or a file)
2. Load data into Pandas DataFrames
3. Restrict to range of interest
4. Resample to days or hours
5. Extract stats (count, min, max, avg and so on)
6. Tabulate and plot

Roadmap / Potential TODO

1. Verify against public Gerrit instances
2. Improve testing
3. Improve setup config
4. Investigate use of Pandas Panels (deprecated) or MultiIndex instead of lots of DataFrames.
5. Add statistics for jobs run (we collect this data, just a case of assembling it)
6. Add in highlighting of out of bounds values in the html report e.g. excessive failure rates in red
7. Other reports/stats requested:
   • how long changes are queued before getting a ci run (may need to query Zuul for this)
   • ci system utilisation/capacity (used total ci hours in the past here, alternative is to aggregate cpu/memory usage stats from the systems)
   • number of Zuul gate queue resets for a change (may need to query Zuul for this)
8. Job duration handling may only work for some Gerrit instances for now, need to determine if better way of gathering that.
9. CI capacity stat relies on job duration data, so this is also unavailable on some Gerrit instances.

License

This project is licensed under the Apache 2.0 license. Please see LICENSE for more info.

Contributing and feature requests

Contributing: You know the drill. Fork it, branch it, change it, commit it, and pull-request it. We are passionate about improving this project, and glad to accept help to make it better. However, keep the following in mind:

• You must sign a Contributor License Agreement first. Contact one of the authors (from Hewlett Packard Enterprise) for details and the CLA.
• We reserve the right to reject changes that we feel do not fit the scope of this project, so for feature additions, please open an issue to discuss your ideas before doing the work.

Feature Requests: If you have a need that is not met by the current implementation, please let us know (via a new issue). This feedback is crucial for us to deliver a useful product. Do not assume we have already thought of everything, because we assure you that is not the case.