Triaging Launchpad project bugs
Our triage process is basically this: make sure that Critical and High bugs are correctly marked.
Critical bugs to be those that need attention before all others. Right now: OOPSes, timeouts, regressions, stakeholder-escalated bugs, operational issues (e.g. build breakage, code issues causing deployment failures, things preventing us detecting other failures such as cronspam, things that should oops but a lack of tooling prevents) and bugs that are dependencies of other critical bugs.
The High bugs list to be around six months deep. Many parts of Canonical are on a six month cycle and fitting in with that is convenient. Some specific sorts of bugs we always treat as high. Right now: A and AA treat accessibility bugs.
We use a quarterly review to shrink the High list if it looks like more than six months of work.
For a full understanding of why we triage bugs and how we came to develop this process, please read our description of the background to our bug triage process.
How to triage
These are the questions we ask when triaging bug reports about Launchpad-project:
Is this a bug in Launchpad-project? If not, move it to the appropriate project (e.g. Ubuntu) and move to the next bug. Note that bugs in lazr.restful, loggerhead etc are bugs in Launchpad-project.
Is this bug on the right subproject? If not, move it to the right sub project.
Is it a duplicate? if there is a duplicate, mark the newer bugs as a duplicate of the older bug.
Is it something we will not do and would not accept a patch to do? If so, mark it as Won't Fix.
Is it an operational request? If yes, convert it to a question.
When are we likely to fix this? Set the importance to show when we'll get to fixing this bug (read more about choosing an importance).
Does the report have enough detail? If we couldn't replicate or otherwise begin work on the bug with the information provided, request further information from the reporter and mark it as Incomplete and move to the next bug. If someone has already asked for more info and the reporter has replied, change the status from Incomplete to Triaged.
set the status to Triaged.
If you're uncertain what importance to give a bug, chat with another engineer. If there's a disagreement, let common sense and courtesy take priority.
Need help? Talk to someone.
All of Launchpad
We use three of Launchpad's bug importances and give each a specific meaning.
Bugs that need to jump the queue. When all is well, we should have no Critical bugs.
Bugs that are likely to get attention in the next six months.
All other bugs.
The importance of a particular bug report reflects the priorities of the Launchpad project. Individuals working on Launchpad may have different priorities. (Read more about selecting bugs to work on)
Any bug marked Critical takes priority over all other bugs.
At present, timeouts, OOPSes (thanks to our zero OOPS policy), security bugs, regressions (including supported-browser issues) and stakeholder escalations are all marked as Critical. Non-security bugs should also be tagged "oops", "regression", etc. so that the reason for their importance is clear. Other types of bug may also be Critical; Francis or Robert will expect you to justify marking any other type of bug as Critical.
If all is well with Launchpad, there should be no Critical bugs.
These are bugs that we believe we will work on in the next six months and A and AA conformance accessibility bugs.
We mark as Low any bug that we recognise as legitimate but that is not scheduled for Canonical staff to fix in the next 6 months. This is not the same as planning not to fix the bug; it means that we don't know when we will fix it, if at all. This includes AAA conformance accessibility bugs.
We do not use Medium or Wishlist. This is primarily to avoid giving false hope to people who are interested in a bug that is neither Critical nor High: if it does not have one of these statuses, we think it is unlikely we will fix it in the next six months.
We tag bugs as part of the triage process. Read the list of Launchpad tags to find out which tags to use.
We do not assign bugs as part of the triage process. Only In progress bugs should be assigned to someone.
Even Critical bugs do not need an assignee, unless they are being worked on. Being at the top of the queue is all we need for Critical bugs to get the attention they require.
Selecting bugs to work on
If you are working on Launchpad in your own time you'll most likely want to fix those bugs that matter to you, regardless of what importance the Launchpad project gives them. That's great and we welcome all bug fixes; we encourage you to look at our page about fixing bugs first.
Members of Canonical's Launchpad team will select bugs depending on whether they're in a maintenance or feature squad.
Generally speaking, squads on feature-rotation will consider the importance of a bug only after filtering for work that applies directly to their current feature.
Maintenance squads, however, will usually be working from the bug database: picking bugs based on their triaged importance. They should look at each importance in order — critical, high, low — and from within that bucket take one of the oldest bugs. Crucially though, there should be no Critical bugs before they start work on High or Low bugs. Engineers should prefer High bugs over Low bugs, but may use their discretion.
Four times a year, we put all of the High bugs back through the triage process. This lets us make sure that all those bugs really should be High and to take account of anything that has changed since they were last triaged.
Beyond these rules a bug is more important than another bug if fixing it will make Launchpad more better than fixing the other bug.
Discretion and a feel for what's in the bug database will help a lot here, as will awareness of our userbase and their needs. One sensible heuristic is to look at five to ten existing High bugs and, if the new bug is less important than all of them, mark it Low as it's probably less important than all existing High bugs.
Engineers have discretion to decide any particular bug should be sorted higher (or lower) than it has been; some change requests are very important to many of our users while still not big enough to need a dedicated feature-squad working on them.
When two engineers disagree, or if someone in the management chain disagrees, common sense and courtesy should be used in resolving the disagreement.