Ubuntu Error Tracker in R
Hadoop
Moving apport and whoopsie to upstart inotify watches, if it's ready
Team views
-proposed view
server-side hooks
"What makes this problem interesting?" section
Package installation failures, kernel oopses, application hangs, and debconf dialogs
Using TPUT to speed up our most common problems calculation
JMX to Graphite
Staging server
Nagios checks
Retracer stability
What do we consider a regression?
Should we email/text/etc if a published binary is showing errors that it purports to fix?
...
Other Error Tracker blueprints for Raring:
https:/
https:/
https:/
https:/
https:/
Blueprint information
- Status:
- Not started
- Approver:
- Steve Langasek
- Priority:
- Undefined
- Drafter:
- Evan
- Direction:
- Needs approval
- Assignee:
- Evan
- Definition:
- New
- Series goal:
- Accepted for raring
- Implementation:
- Unknown
- Milestone target:
- ubuntu-13.04
- Started by
- Completed by
Whiteboard
Work Items
Work items for ubuntu-
[pitti] Figure out how to create an ARM retracer: DONE
[pitti] Support gdb-multiarch and multiple architectures in apport-
[pitti] Improve the retracer charm to set up ARM retracers: DONE
[pitti] Turn off crash reporting to Launchpad in 13.04 - only have them go to daisy / errors (nacked by desktop team, needs "request LP bug" feature from per-problem hooks): POSTPONED
[pitti] Create tool to generate predictable binary and Python error reports that we can use for end to end system testing: DONE
Work items:
Deploy Hadoop on Cassandra (https:/
Create a higher level interface to creating Hadoop jobs using Pig or Hive: BLOCKED
Support errors from debconf dialogs: POSTPONED
Support errors from hangs using compiz: POSTPONED
Support errors from kernel OOPSes: DONE
Support errors from package installation failures: DONE
Support errors from GPU hangs: POSTPONED
Support errors from suspend/resume failures: POSTPONED
Support Ubuntu installation failures: POSTPONED
Patch the Mozilla error reporting mechanism to also send data to daisy.ubuntu.com (bug #1064395): POSTPONED
[brian-murray] Import developer to package and developer to team mappings: DONE
[brian-murray] filter the default view on errors.ubuntu.com to problems the developer is responsible for: DONE
Backport Raring whoopsie to Precise and Quantal: TODO
Provide an automatic error reporting option for apport-cli and apport-gtk: TODO
Set a field when automatic error reporting is enabled, so we can track its usage: TODO
[mpt] UI design for enabling automatic error reporting on the desktop UI <https:/
[mpt] UI design for disabling automatic error reporting in the System Settings panel <https:/
Add an automatic error reporting opt-in debian-installer component: POSTPONED
Add automatic error reporting to cloud-init: TODO
Work with the Juju team to define and implement constraints that enable automatic error reporting: TODO
Work with the Landscape team to get an automatic error reporting option on the roadmap: TODO
Move whoopsie from long-running to using upstart's inotify events: BLOCKED
Implement TPUT to speed up our most common problems calculation: TODO
[ev] Fix the race conditions in the juju deployment system: DONE
Define initial set of alert conditions: DONE
Implement alert condition support: TODO
Implement web UI for server-side hooks: TODO
Implement ACL for server-side hooks: TODO
Implement expiry system for server-side hooks: TODO
Implement the Daisy side of server-side hooks: TODO
Move retracers to Squid proxy for package caching (RT 57114): DONE
make a video regarding how to use errors.ubuntu.com: POSTPONED
Nagios checks for the retracers: TODO
Nagios checks for submitting each problem type (that is simulated or gets undone): TODO
Nagios checks for the errors.ubuntu.com API: TODO
Engage with Chrome and Mozilla upstream to get them access to the data for their respective applications: TODO
Implement expanders on the problem page which show the debdiff between the first seen version and the version prior: TODO
Investigate the accuracy of detecting errors originating in library dependencies when the library version is present on all instances (discussed at sprint, not reliable): DONE
Create a basic "what's interesting about this problem?" section: TODO
Feed package fields into the "what's interesting?" section (low on disk space, arch, date, etc): POSTPONED
Provide realtime updates to errors.ubuntu.com using txlongpoll: POSTPONED
[mpt] UI design for optionally providing a description of an error that just occurred (bug 1084624, won't fix): DONE
implement sending the description of an error that just occurred: TODO
[mpt] UI design for presenting errors that do not need immediate explanation together with the next one that does <https:/
Implement sending errors that do not need immediate attention with the next report that does: INPROGRESS
[mpt] Work with mathematicians in the community to develop a more accurate average errors per calendar day calculation: INPROGRESS
Finish implementing -proposed view to facilitate SRU work: TODO
Decide whether we can set /proc/sys/
Work with Stefan and the kernel team to establish whether we can send kernel crash reports, possibly in a reduced-size format (to limit memory impact): INPROGRESS
[ev] Work with Stefan and the kernel team to establish whether data on kernel warnings will be useful: INPROGRESS
[ev] Discuss privacy of SHA-512 hashed MAC addresses with the security team: INPROGRESS