kitchen: monitor failures in recipe engine |
|||||||||||||||
Issue description
if a recipe engine, not the user recipe, fails, kitchen must mark the build as an internal failure. We must get notified when this happens.
proposal:
- implement as a counter metric for a builds with fields "status" ("success", "failure"), "failed component" ("", "kitchen", "recipe_engine", "recipe", "logdog"?).
- implement on alert on positive number of non-recipe failures
Re counter: there are many instances of kitchen, so either have to
- introduce new field "bot_id". Won't work if we have builds too often.
- introduce new field "swarming_run_id" and use autogen to collapse swarming_run_id fields in a precomtutation
,
Apr 5 2017
,
Apr 27 2017
,
May 4 2017
Assigned to M1-S1
,
May 11 2017
,
May 11 2017
,
Jun 7 2017
,
Jun 9 2017
,
Jun 20 2017
,
Nov 8 2017
,
Jan 31 2018
,
Jan 31 2018
,
Feb 13 2018
,
Feb 15 2018
,
Mar 22 2018
marking as dup of bug 721576 because if bug 721576 is fixed, the build will infra failure. We have monitoring for infra failures |
|||||||||||||||
►
Sign in to add a comment |
|||||||||||||||
Comment 1 by no...@chromium.org
, Feb 23 2017Labels: luci