Architecture decisions, post-mortems, infrastructure write-ups. Honest about what worked and what didn't. Subscribe via RSS or email.
Three years of evolution from datacenter IPs to residential rotation to a hybrid pool. The cost decisions, the mistakes, the load-balancer config we landed on.
Tomás Vega · 02 May 2026A 47-minute outage in our task queue. The contributing factors (bad assumption + cascading retry storm), the recovery playbook, and what we changed in the runbook.
Anita Sharma · 28 Apr 2026We started with 9 services. We're now at 1 monolith + 1 worker pool. Why we made the change, what we lost, what we gained, and the migration plan.
Tomás Vega · 22 Apr 2026Maps changes its DOM ~once a quarter. Each break has a fingerprint. We catalog them, automate detection, and roll patches without you noticing. Here's the system.
Naomi Okafor · 18 Apr 2026The dashboard home page was slow. We profiled, found the culprit (an N+1 in the activity feed), and shipped a fix that improved p95 by 11×. The full investigation.
Naomi Okafor · 11 Apr 2026SOC 2 audits get a bad rap. With the right tooling (Drata) and a couple of quiet process changes, we kept our weekly release cadence and still got audited cleanly.
Anita Sharma · 04 Apr 2026~2 engineering posts per month. RSS feed available too.