Notes from the engineering bench.
Page 3 of 5
Blog posts — page 3
API versioning strategies that don't hurt later
Three ways to version an API, one way Stripe does it, and one trap that turns every v2 into a reimplementation project. Here's how to pick a strategy that survives three years of shipping without forcing a rewrite on every integrator.
Churn analytics: detecting, preventing, and recovering from SaaS churn
Most SaaS churn reports are lagging indicators dressed up as insight. Here's how to track the signals that actually give lead time — and the recovery playbooks that justify the engineering work.
Job queues in Node.js: BullMQ, Inngest, Trigger.dev, or build your own?
Every Node.js backend eventually needs a queue. The choice between BullMQ, Inngest, Trigger.dev, a Postgres table, or SQS shapes your ops surface for years. Here's an honest comparison with the tradeoffs that matter at 100, 10K, and 10M jobs a day.
Real-time coordination for marketplaces: Socket.IO patterns that scale
Marketplaces live or die on coordination — two users need to see the same booking state, message thread, and availability window within the same second. Here are the Socket.IO patterns we actually ship, and how they behave once the server count crosses one.
Multi-tenant SaaS architecture: schemas, isolation, and scaling patterns
Shared schema, separate schema, or database-per-tenant? How to pick, how to enforce isolation at the database layer, and how to stop one greedy tenant from ruining everyone else's day.
Observability for SaaS: metrics, logs, traces, and the tools that matter
Observability is where SaaS teams quietly overspend. Here's a working guide to the three pillars, the tools worth using in 2026, and the cost-control decisions — sampling, cardinality, retention — that separate a $200/month bill from a $20,000 one.
Handling refunds, disputes, and chargebacks in SaaS
SaaS is uniquely exposed to chargebacks because the product is intangible and subscriptions auto-renew. Friendly fraud is now up to 70% of card losses. Here is the refund policy, evidence package, and dunning discipline that keep dispute ratios under 0.5%.
Two-sided marketplaces in 2026: the playbook for solving chicken-and-egg
Every two-sided marketplace starts with the same dead start: no supply means no demand, and no demand means no supply. The teams that solved it — Airbnb, Uber, DoorDash, and the less famous ones that copied the pattern — did not find a trick. They picked one side, faked it until it was real, stayed geographically dense, and only expanded when the first market stopped needing their attention.
Webhook security: HMAC verification, replay protection, and idempotency
Webhook endpoints are the most exposed public surface in most SaaS products. Here's how to verify HMAC signatures correctly, block replays, and handle duplicate deliveries without double-processing a single event.
Next.js SEO in 2026: metadata API, sitemaps, and Core Web Vitals
SEO in 2026 is less about meta-tag trivia and more about engineering discipline. Next.js ships most of what you need — but only if you wire it up correctly. Here's the playbook: metadata API, dynamic sitemaps, structured data, and Core Web Vitals targets you can actually ship against.