Process Review - 2024.10.17
Planning/Pre-development
-
Discuss possible ripple effects of changes - document in QA field
-
Discuss/document if coordinated releases are needed with Atlantis/Utopia - clear release plan
-
Identify new components or tricky components to change
Testing/Release
-
At handoff from Dev to QA discuss possible ripple effects of changes - document in QA field
-
When QA going through testing spreadsheet mark items that have an Atlantis/Utopia dependency
-
Default of both Dev and QA testing on Local
-
Try to avoid running Atlantis Frontend & Utopia both locally - test components independently
-
If needed, coordinate with QA and stage your branch before merging into the main branch
-
If we have multiple people needing to stage things at the same time, will spin up resources like Nirvana
-
Post-release
-
Handling jump balls from #product-stability - On-call Devs as triage
-
CloudOps (often Emily) reaches out to QA if it seems like it is code related
-
QA determine if it is a regression vs. an old bug
-
On-call Devs should also watch #product-stability for bugs showing up there
-
QA follow up with the broader team on #qa-dev or #war-room depending upon the severity and tag the person/team it is relevant for, if known
-
On-call Devs should watch #qa-dev for newly discovered regressions that don’t have a clear team/person to fix and/or need to be researched on the cause
-
-
Once the issue is determined, unless it is a quick fix for the On-call Dev then they pass it on to the person/team from the story that triggered the bug
-
Once severity and scope of fix determined, consult with team and/or broader group on #qa-dev to decide path forward - Rollback or partial rollback / Hotfix / Fix with next regular release
-
David/PMs communicate with the company in appropriate channels about bug and hotfix or rollback
-
Plan ahead on rollback plans and/or feature flags (for bigger, more complicated items)