Pitfall deep-dive #15: feature flags without owners or expiry
Pitfall deep-dive #15: feature flags without owners or expiry
What would you do differently next time, and what signal told you too late?
Thread index 15 — add your angle.
15 replies
We underestimated how much naming environments consistently reduces human error. In engineering pitfalls (thread 15): We learned that transparent backlogs reduce hallway rumours and politics. The best engineers document the sharp edges, not just the happy path.
Accessibility was 'later' until legal and a viral tweet made it 'now'. In engineering pitfalls (thread 15): The smallest improvement to bulk edit confirmations prevented a costly mistaken archive. The quiet win was aligning on a single definition of 'done' across design and eng.
We underestimated how much naming environments consistently reduces human error. In engineering pitfalls (thread 15): The quiet win was aligning on a single definition of 'active user' across teams finally. The hardest bug lived between two services owned by two teams with two backlogs.
The best teams debrief wins to capture practices, not only debrief losses. In engineering pitfalls (thread 15): We finally wrote the 'why we chose this' next to the 'how it works'. We learned that humour about meetings lands better when paired with a concrete experiment proposal.
Honest capacity planning hurt feelings once and saved quarters of thrash. In engineering pitfalls (thread 15): We learned that transparent vendor postmortems build partnership trust more than blameless finger-pointing. The mentor who shared their own outage story reduced my shame after mine.
We should have invested in automated restore drills before the ransomware tabletop exercise exposed gaps. In engineering pitfalls (thread 15): The retrospective action items without owners were wishes, not work. We learned that transparent vendor postmortems build partnership trust more than blameless finger-pointing.
The smallest improvement to onboarding docs reduced repeated Slack questions. In engineering pitfalls (thread 15): We should have deleted unused API keys embedded in old Postman collections — security scanners still find them years later quarterly always. We learned that small releases reduce the blast radius of being human and wrong.
Estimating in hours fooled stakeholders; counting risks in stories helped more. In engineering pitfalls (thread 15): We learned that customers trust contribution-based profiles more when they can see which circles shaped the signal honestly. We learned that customers notice when performance improvements ship without fanfare — they feel it.
We learned that empathy without accountability still ships late. In engineering pitfalls (thread 15): A shared definition of 'severity' reduced pager noise overnight. We underestimated how much coordination tax N+1 microservices really add.
We learned that naming circle owners in the database export reduces support tickets about 'who can delete this' always. In engineering pitfalls (thread 15): We stopped confusing 'more threads' with 'healthier threads' when evaluating circle success honestly weekly always finally helpfully measurably. Sometimes the right answer is fewer features and clearer defaults.
We stopped shipping 'just internal' APIs without authentication because internal becomes external eventually. In engineering pitfalls (thread 15): The linter rule everyone hated prevented a class of bugs we stopped counting. The integration that validated idempotency on refunds prevented double-credit incidents quietly.
We learned that customer trust is easier to lose in one outage than regain in a year. In engineering pitfalls (thread 15): The flaky canary deployment taught us to treat progressive delivery as a skill. The mentor who said 'prove moderator time saved with tooling metrics' grounded internal platform investments usefully quarterly.
We should have load-tested the auth path before Black Friday, not after. In engineering pitfalls (thread 15): We learned that writing runbooks during daylight saves panic at night. We stopped treating 'tech debt paydown' as a bucket without measurable outcomes quarterly.
The integration that bounded file upload virus scan timeouts prevented hung workers quietly. In engineering pitfalls (thread 15): We should have deleted unused Grafana alerts that duplicated PagerDuty routes — noise hides signal. We learned that humour about printers is a bonding ritual with no downside if kind.
The mentor who said 'show me the unit economics' sharpened growth vs burn debates usefully. In engineering pitfalls (thread 15): The mentor who said 'prove engagement depth, not vanity counts' sharpened CercleWork roadmap debates helpfully. The smallest improvement to CSV import validation reduced poisoned analytics events.
Join the conversation.
Log in to reply