FIELD NOTES FROM THE GRAVEYARD.
Engineering deep-dives, triage guides and live-ops war stories from the team building Tombstack.
How we cut crash → dashboard latency to under 40ms
A deep dive into the ingest pipeline that lets a studio see a crash spike before the player even rage-quits — careful batching, fingerprinting, and async symbolication.
Reading a Tombstack: anatomy of a good crash report
Stack trace, breadcrumbs, device state — what each part of the autopsy tells you, and the order to read them in when you're triaging under fire.
Funnels that don't lie: instrumenting your onboarding
Most onboarding funnels measure the wrong thing. Here's how to fire events that map to real player intent, and how crashes silently poison your conversion.
Surviving launch day: a live-ops runbook
CCU is climbing, a region just hit capacity, and a crash spike just paged you. What to watch, in what order, and when to roll back.
Symbolicating crashes across PC, console and mobile
Three platforms, three symbol formats, one readable stack trace. How we normalize PDB, dSYM and breakpad into a single grave.