Route by service ownership and severity so the right person wakes up — and nobody else.
The best way to reduce alert fatigue is not to raise thresholds until the noise stops — it is to make sure each alert reaches the one person who can act on it. Today we are shipping ownership-aware routing to do exactly that.
Route on the data's labels
Because Helix alerts carry the labels of the telemetry that triggered them, they already know which service fired. Routing now maps that service to its owning team automatically, so a database alert never wakes a frontend engineer again.
alert "error spike"
| when rate(errors) by service > 0.05 for 5m
| route to team_of(service)
| page if severity == "critical" else ticketTwo severities, two behaviors
- Critical, customer-facing alerts page a human immediately.
- Degraded-but-not-down alerts open a ticket instead of ringing a phone.
Available now
Ownership routing is live for every team. Point your services at their owners once, and the pager gets quieter on its own.