Ensure all future work with cross-team scope is flagged and reported to relevant groups for approval and awareness

Summary

We encountered an incident where we modified a query which was powering another team's feature. Our query overloaded the database and broke this feature. We should ensure this does not happen again by modifying our processes and communications to ensure we flag work proactively to any involved teams.

Implementation guide

  • When raising new MRs which involve changes to the codebase beyond our own areas, we should proactively reach out to affected teams even if CODEOWNERS or similar automations do not trigger this.
  • We should make additional effort to find the right people to inform about our changes, even if it's only for visibility and not code review.
  • If we're uncertain about ownership (eg. CODEOWNERS or similar does not identify an owner), engineers should explicitly flag this and ask for help identifying DRIs who can be involved.
  • We should not merge any new work which touches on areas beyond immediate grouporganization scope/ownership until we have found the relevant owner and ensured they are aware of our impact/plan.
Edited by Matt Andrews