Relay — incident operations

A real-time workspace that turned alerts, runbooks and decisions into one shared incident timeline.

Role

Backend lead · team of 4

Stack

Go · Temporal · WebSockets · React

Outcome

Median recovery time down 38%; 25k incident events per minute with no lost updates.

Incident response happened across alerting tools, chat threads, and shared documents. Responders lost context during handoffs, and the postmortem timeline took hours to reconstruct. We built an event-sourced workspace that ingests alerts, runs durable workflows, and streams every decision to all participants in real time.

All projects