Projects
I picked up reviews on two infrastructure PRs from Leo this session: "Add terraform code for new component new-typey-sale-denormalizer-consumer" (#3788 on terraform-services) and "Add new typey-sale-denormalizer-consumer component" (#1445 on catapult). Both got change requests. I also reviewed Vicky's PR for LRX-31817, which updates the Sale History Actions UBE in monocle. On my own work, I made another pass at the sortable column headers PR (#15185 on monocle for LRX-31809), addressing review feedback and doing a light refactoring pass to tighten things up.
Team
Self review is finally done and submitted in Workday. Mentioned yesterday I was not letting it slip. It did not slip.
Sprint Retro
What Went Well
Closing both typey backend tickets (LRX-32360 and LRX-32369) alongside the graphter sort ticket (LRX-32361) mid-sprint turned out to be the right call. It unblocked all seven FE sort tickets (LRX-31809 through LRX-31816) in one go, and they moved to In Review together as a coherent batch rather than trickling in one at a time. On the personal project side (rules-as-written), throughput was unusually strong: 15 or more PRs merged in under two weeks, spanning CI consolidation, SSR fixes, accessibility remediation, a CMS migration, and new features. That kind of sustained velocity on a side project is worth noting and repeating.
What Didn't Go Well
LRX-34876, the French translation bug for "Cash" and "Gift Card", is still open. It is a straightforward localization fix, not a hard problem, but it carries real user-facing weight in a Quebec context and it quietly got buried under the sort feature work. LRX-36507, the external PR review for custom tax rate precedence, is also still sitting open. External reviews have a way of stalling without anyone noticing until a sprint retro surfaces them. Meanwhile, all seven FE sort tickets landed In Review but none closed. The backend was fully ready by March 17, so the bottleneck has shifted squarely to the frontend review and merge side.
What I Learned
Landing LRX-32360 and LRX-32369 on the same day was a deliberate sequencing choice, and it paid off. When backend work arrives as a unit, the frontend team can move as a unit rather than waiting on piecemeal approvals. The SSR edge cases that surfaced after the Party Tracker MVP shipped on rules-as-written (issues #75 and #76) were a useful reminder that correctness at the Netlify function boundary is not something you discover by accident in production. That check belongs before a feature PR closes, not after.
What to Do Differently Next Sprint
External reviews like LRX-36507 need a hard timebox. If there is no response within two business days, the right move is to escalate rather than wait. LRX-34876 needs an explicit slot on the board next sprint with an owner and a priority signal rather than floating open indefinitely. On the rules-as-written side, I am setting a personal rule: no new feature PRs on a route until its SSR path has been smoke-tested in a deploy preview.