Tool to crawl a Jenkins site using a build url and producing a SVG output to render the build flow.
The SVG graph displays: - A summary box with:
- the total elapsed time
- the cumulated duration for each build
- a throughput (duration/elapsed)
- number of builds
- Black arrows to render upstream and downstream relation
- Orange arrows to render downstream only relation
- Build with a blue/yellow/red/gray box for Success/Unstable/Failed/Aborted build status
Build information are stored in a local sqlite database. The database is used as a cache to not fetch twice a build page but also to get information using plain SQL:sqlite3 ~/jenkviz.db -- Slowest jobs sqlite> SELECT name, SUM(duration_s), MAX(duration_s), AVG(duration_s), COUNT(1) FROM build GROUP BY name ORDER BY SUM(duration_s) DESC LIMIT 10; -- Slave load sqlite> SELECT host, SUM(duration_s) FROM build GROUP BY host ORDER BY SUM(duration_s) DESC LIMIT 10;
jenkviz [--version] [--logfile=LOGFILE] [--database=DATABASE] COMMAND [OPTIONS] ARGUMENTS
crawl [--direct|--reverse|--explore] [--output SVG_FILE] JENKINS_BUILD_URL
The --direct option shows only downstream and upstream relation, removing downstream only link.
The --reverse option crawl backward using upstream builds.
The --explore option to keep downstream builds that have upstream build out of the scope of the origin build (the upstream build is not a descendant of the root build)
jenkviz crawl http://jenkins.site/jenkviz/job_name/42/
Jenkviz try to find downstream/upstream build using the web page and the REST API to work around the JENKINS-6211 bug.
On Debian/Ubuntu:sudo aptitude install graphviz sudo easy_install jenkviz
Jenkviz is currently hosted at github.
Feature requests and bug reports can be made here: