Modern web applications can quickly develop many moving parts. As they grow in traffic, infrastructure, services and features, keeping track of errors and communicating about them becomes more challenging.
In this talk we present Triage, a homegrown tool at 99designs for aggregating and managing errors in a live production system. Although written in Python, Triage is designed to ingest errors in a language agnostic way. We describe our path from simpler logging techniques which we outgrew, the new workflows around error management which it allows, and some challenges we faced along the way.