This is a hard topic to discuss, as there is no black-or-white answer. I am OK with this protocol. It is critical to leave a note somewhere so that the merged PR can be verified by an expert soon when he/she becomes available.
A long-term solution is to have more committers who are familiar with the related modules. It means we need to grow the developer community. Chen Li On Wed, Sep 9, 2026 at 8:57 AM Ali Risheh <[email protected]> wrote: > Hi everyone, > > I’d like to initiate a discussion regarding a workflow we used last week to > unblock an urgent PR when standard code reviewers were unavailable. We want > to gather community feedback on whether this should be an accepted practice > and, if so, how to formalize it. > > *Context & Challenge* Last week, we had a time-sensitive PR that was > blocked due to two factors: > > 1. > > *Domain Expertise:* The PR touched Kubernetes and specific > infrastructure code that available team members did not feel confident > reviewing line-by-line. > 2. > > *Availability:* The subject matter experts familiar with this area were > unavailable. > > *Action Taken* To avoid delaying development, we held an ad-hoc sync with > two committers and one IPMC member. Instead of performing a standard > line-by-line code review, we reviewed and validated the high-level design. > Once consensus on the design was reached, we approved and merged the PR. > > *Discussion Points / Questions* > > 1. > > *Policy Approval:* Should we formalize design-level consensus as an > acceptable alternative for urgent PRs? If not, what process should we > use > to unblock development when domain reviewers are unavailable? > 2. > > *Quorum & Governance Requirements:* If this protocol is acceptable, what > minimum criteria should we establish? (For reference, this instance > involved a PPMC-authored PR with design approval from two committers and > one IPMC member. Would a lower threshold, such as one committer > approval, > be sufficient?) > > Please share your thoughts so we can determine the best path forward. > > Best regards, > > Ali >
