Skip to content

Issue triage process

Seth Ladd edited this page Sep 12, 2016 · 4 revisions

(Meta: not trying to make a laborious, or super-detailed, triage process. Use your best judgement.)

Status: draft

Objective: Developers contributing to Flutter can easily and quickly choose impactful issues.

Strategy:

  • Ensure each issue is actionable.
  • Ensure each issue is prioritized.
  • Close issues with the assumption we can always reopen.

When reviewing an issue, ask yourself the following questions:

  • Is this issue still valid?
  • Is it a dupe?
  • Is it high or low priority? (affecting partners/goals)
  • Do we have enough info to take action?
  • Do we know how to tell if it's done?
    • Sometimes, issues might be valid, but are very very high level (e.g. "It's slow"). We prefer issues that we can test if we've fixed the problem.
  • Do we need more info?
  • Do we envision taking action on it in the next 6 months?

Then, make the following changes to the issue:

  • Engage with the customer/user, empathize with them, get more info to make the issue actionable.
  • Close an issue if we don't intend to work on the issue ever.
  • If we need more info, add "waiting for customer response" label.
    • If we haven't heard back in two weeks, close the issue.
  • Put the issue into the correct milestone, indicating its priority.

End result:

  • Our engineers have a short list of impactful, actionable issues to choose from.
  • Customers feel heard.

Clone this wiki locally