Classify tempest-devstack failures using ElasticSearch
Go to file
Joe Gordon b0b024e080 Wait till all log files are in ElasticSearch
We push each log file into ElasticSearch search separately, so knowing one
file is there doesn't guarantee a different file from the same job is ready.
So instead of checking for just console.html use facets to make sure all
the relevant log files are ready.
2013-09-18 13:55:37 -07:00
tests Wait till all log files are in ElasticSearch 2013-09-18 13:55:37 -07:00
.gitignore Add dot files 2013-09-13 00:49:28 -07:00
.testr.conf Add dot files 2013-09-13 00:49:28 -07:00
bot.py Add irc bot support to elasticRecheck 2013-09-17 00:27:14 -04:00
elasticRecheck.conf Add irc bot support to elasticRecheck 2013-09-17 00:27:14 -04:00
elasticRecheck.py Wait till all log files are in ElasticSearch 2013-09-18 13:55:37 -07:00
LICENSE First commit, including proof of concept ElasticSearch tool 2013-09-12 16:54:27 -07:00
proof_of_concept.py Add reqs file 2013-09-12 17:01:32 -07:00
queries.json add query for bug 1226791 2013-09-17 13:00:20 -07:00
README.md Add irc bot support to elasticRecheck 2013-09-17 00:27:14 -04:00
recheckwatchbot.yaml Add irc bot support to elasticRecheck 2013-09-17 00:27:14 -04:00
requirements.txt Add irc bot support to elasticRecheck 2013-09-17 00:27:14 -04:00
tox.ini Add tox.ini file for flake8 ignores 2013-09-18 13:47:18 -07:00

elasticRecheck

Classify tempest-devstack failures using a list of elastic search queries.

Idea

When a tempest job failure is detected, by monitoring gerrit (using gerritlib), a collection of logstash queries will be run on the failed job to detect what the bug was.

Eventually this can be tied into the rechecker tool and launchpad

Future Work

  • Pull in list of queries from a more flexible source, so a commit isn't needed to update each time
  • Turn into a server app
  • Make unit tests robust and not need internet

Dependencies

  • gerritlib
  • pyelasticsearch