CI Bug Log issueshttps://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues2020-01-02T12:26:20Zhttps://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/2Compare: Consider it a failure if 50% of tests were not seen in the new runcfg2020-01-02T12:26:20ZMartin RoukalaCompare: Consider it a failure if 50% of tests were not seen in the new runcfgThe equivalent for hosts was already done in 5627c587edcd3543eb2a8ed5837ca32e8f446203.The equivalent for hosts was already done in 5627c587edcd3543eb2a8ed5837ca32e8f446203.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/5Compare: List some statistics per host (#tests, execution time, pass rate, ...)2019-01-25T12:43:38ZMartin RoukalaCompare: List some statistics per host (#tests, execution time, pass rate, ...)This will help get an overview of the run :)This will help get an overview of the run :)https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/6Compare: Warn when the execution time of a test/runconfig grows too much2019-01-25T12:44:20ZMartin RoukalaCompare: Warn when the execution time of a test/runconfig grows too muchThis should allow us to prevent drastic performance regressions.
This may loosely depends on #5.This should allow us to prevent drastic performance regressions.
This may loosely depends on #5.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/7Run import: Give the list of the hosts that were supposed to run tests2019-01-25T12:45:16ZMartin RoukalaRun import: Give the list of the hosts that were supposed to run testsThis will allow to enhance #2, but I heard it is not that easy :) Some work has however already been done on the CI side to start keeping track of this.This will allow to enhance #2, but I heard it is not that easy :) Some work has however already been done on the CI side to start keeping track of this.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/11Send duallistbox patches upstream2019-01-25T12:53:24ZMartin RoukalaSend duallistbox patches upstreamTo avoid having to rebase our changes over and over again, we should land as many of our changes into duallistbox :)To avoid having to rebase our changes over and over again, we should land as many of our changes into duallistbox :)https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/13Periodically analyze issues to find out which ones can be closed2019-08-20T05:04:42ZMartin RoukalaPeriodically analyze issues to find out which ones can be closedThis will help us with bug tracking, and we could post a comment when actions need to be taken!This will help us with bug tracking, and we could post a comment when actions need to be taken!https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/15Allow creating dashboards in the metrics view2019-02-15T05:13:19ZMartin RoukalaAllow creating dashboards in the metrics viewWe could create a way to split a page to load multiple metrics which would allow creating dashboards per platform for bugs, passrate, etc...
This should be easier when #9 is implemented, since the query will be stored in a single variab...We could create a way to split a page to load multiple metrics which would allow creating dashboards per platform for bugs, passrate, etc...
This should be easier when #9 is implemented, since the query will be stored in a single variable :)https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/18Sorting by "last updated" in ci-bug-log doesn't work.2019-01-31T14:02:08ZRadosławSorting by "last updated" in ci-bug-log doesn't work.When sorting by last updated we end up with all entries that start with 1 (1, 10, 100 days) followed by 2 and others.
Instead we should have 1,2,10 etc.When sorting by last updated we end up with all entries that start with 1 (1, 10, 100 days) followed by 2 and others.
Instead we should have 1,2,10 etc.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/19Try to match unknown failures associated with a machine when we update the ta...2024-01-30T16:18:00ZMartin RoukalaTry to match unknown failures associated with a machine when we update the tags of that machine**What?**
When we change machine's tags (especially add a new one) we should try to match all unknown failures associated with that machine with the updated set of filters (some filters are tied to tags).
**Why?**
Sometime we are late...**What?**
When we change machine's tags (especially add a new one) we should try to match all unknown failures associated with that machine with the updated set of filters (some filters are tied to tags).
**Why?**
Sometime we are late with adding tags after changing machine configuration which leaves us with a bunch of unknown failures. Currently there is no way of matching them in automated way after updating the tags.1.0https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/21Provide an occurrence-rate point cloud in the issue/filter history's page2019-04-23T11:46:56ZMartin RoukalaProvide an occurrence-rate point cloud in the issue/filter history's pageThis will make it easier to show the reproduction rate and make it easier for developers to understand what is going on without having to do the math themselves.This will make it easier to show the reproduction rate and make it easier for developers to understand what is going on without having to do the math themselves.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/23Create a command line interface for interacting with cibuglog2019-02-19T14:09:31ZMartin RoukalaCreate a command line interface for interacting with cibuglogThe goal would be to get rid of the add_* scripts. This will need to work for both the development vs container mode.The goal would be to get rid of the add_* scripts. This will need to work for both the development vs container mode.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/24Add a view to show how many tests per week we execute2019-02-28T14:16:44ZMartin RoukalaAdd a view to show how many tests per week we executeThis will be useful to keep track of our execution rate and try to improve that :)This will be useful to keep track of our execution rate and try to improve that :)https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/31Write a end-to-end test for importing results2024-01-30T16:21:40ZMartin RoukalaWrite a end-to-end test for importing resultsI commited a regression to the import process and it took almost a week to realise and fix the issue. Let's prevent this from happening by writing more complex tests that verify the entirety of the import process.I commited a regression to the import process and it took almost a week to realise and fix the issue. Let's prevent this from happening by writing more complex tests that verify the entirety of the import process.1.0https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/32Flaw on sorting by priority on metrics2019-08-10T08:57:39ZKarol KrolFlaw on sorting by priority on metricsSorting by priority on Open Bugs view doesn't work correctly.Sorting by priority on Open Bugs view doesn't work correctly.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/34Backlink from bugzilla to cibuglog issue (public copy)2024-01-30T16:22:37ZDaniel VetterBacklink from bugzilla to cibuglog issue (public copy)Instead of one-way info dumps by cibuglog into bugzilla, would be nice if we could backlink for all the additional information on bug occurence.
Per @mupuf this needs public cibuglog instance or a better scraper.Instead of one-way info dumps by cibuglog into bugzilla, would be nice if we could backlink for all the additional information on bug occurence.
Per @mupuf this needs public cibuglog instance or a better scraper.1.0https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/35Button to expand issue_id and ifa_id queries2019-05-02T05:50:37ZDaniel VetterButton to expand issue_id and ifa_id queriesFor analyzing bugs it's often useful to slightly change the issue query, e.g. add/remove machines/testcases or slightly change the regexes. Currently the only way to do that is manually reconstruct the query the issue/active filter uses,...For analyzing bugs it's often useful to slightly change the issue query, e.g. add/remove machines/testcases or slightly change the regexes. Currently the only way to do that is manually reconstruct the query the issue/active filter uses, and then go from there.
Much more convenient, and much easier to discover the power of cibuglog queries would be if issue_id=$id and ifa_id=$id would have a button to expand them to their full query, so you can directly go and edit stuff.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/36Do not compress as much the PASS -> NEW FAILURE2019-11-14T12:41:33ZMartin RoukalaDo not compress as much the PASS -> NEW FAILUREThis would make reports easier to read without having to go to the issue page. Potentially, we just want to set a threshold instead of an on/off but instead having a threshold.
Known failures do not matter as much and could be compress ...This would make reports easier to read without having to go to the issue page. Potentially, we just want to set a threshold instead of an on/off but instead having a threshold.
Known failures do not matter as much and could be compress immediately.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/42Comparison: Try to display the reproduction rate of known issues2019-08-20T04:45:27ZMartin RoukalaComparison: Try to display the reproduction rate of known issuesWhen seeing a pass -> fail [bug] or crash [bug] -> pass, the reproduction rate of the issue would help give context to developers to understand if their changes are likely to have had an impact on this change, or if this was just flakine...When seeing a pass -> fail [bug] or crash [bug] -> pass, the reproduction rate of the issue would help give context to developers to understand if their changes are likely to have had an impact on this change, or if this was just flakiness (especially in the light of a patch series going through multiple revisions).
Suggested-by: Dhinakaran Pandiyanhttps://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/45Allow replacements in url generation in add_run.py2020-01-03T11:57:27ZPetri LatvalaAllow replacements in url generation in add_run.pycibuglog constructs links to the test results based on the test name and patterns given to add_run.py.
i915 gfx CI hosts test results for an IGT test of name "igt@foo/bar@quz" at URL "igt@foo_bar@quz.html" which is impossible to represe...cibuglog constructs links to the test results based on the test name and patterns given to add_run.py.
i915 gfx CI hosts test results for an IGT test of name "igt@foo/bar@quz" at URL "igt@foo_bar@quz.html" which is impossible to represent with just the url pattern {test}, as {test}.html becomes igt@foo/bar@quz.html. Allow for replacement fields in the pattern so we can properly link to all results.https://gitlab.freedesktop.org/gfx-ci/cibuglog/-/issues/48Open bugs: allow de-duplicating user profiles2019-06-11T12:08:08ZMartin RoukalaOpen bugs: allow de-duplicating user profilesDevelopers have one account per trackers. This makes looking the developer statistics a little difficult because a developer will be seen there multiple times (one per account).
Let's make this view use the profile, and enable people to...Developers have one account per trackers. This makes looking the developer statistics a little difficult because a developer will be seen there multiple times (one per account).
Let's make this view use the profile, and enable people to de-duplicate them (associate all the accounts of a person under the same person).