Quickstart
Your first context-briefed alert in under 10 minutes — one happy path, no theory.
#1. Install the CLI
$ npm install -g @relay/cli
+ @relay/cli@3.2.0 — installed in 4s#2. Connect your alert source
Relay sits between your alerting and your team. Point any source at it — Prometheus, Datadog, CloudWatch, or plain webhooks:
$ relay connect prometheus --url https://prom.internal:9090
✓ Connected. 214 alert rules discovered.Note: Migrating from PagerDuty? The importer brings rotations, escalations, and history in one command.
#3. Attach context to a service
This is the step that ends 2am archaeology — tell Relay where a service's runbook and ownership live:
$ relay service set api-gateway \
--runbook ./runbooks/gateway-latency.md \
--owner team:gateway
✓ api-gateway briefed. Next alert arrives with context attached.#4. Fire a test alert
$ relay test api-gateway --severity 2
→ paging @maya (primary, gateway rotation)
→ runbook attached: gateway-latency.md
→ similar: #3971 resolved in 11m — fix linked
✓ Test incident #4129 created — check Slack.That's the whole loop: alert in, context attached, right person paged. From here:
- Route by ownership — page the engineer who shipped the change
- Escalation policies — escalate only when it's earned
- Auto-drafted postmortems — the timeline writes itself