Alerts & incidents
Fewer pages. Better incidents.
Servbyte alerts you when something actually matters, groups related noise into a single incident, and keeps a full timeline so you can review — and prevent — recurrences.
Threshold alerting
Static rules on any metric or service state — CPU %, memory, disk, restarts, request errors. Simple to write, easy to reason about.
Anomaly detection
Baselines per host and per service. Alert when today's behavior diverges from the last two weeks, without you tuning a number.
Noise reduction & grouping
One bad host produces one incident, not fifty pages. Related alerts fold into a single incident automatically.
Routing to your on-call tools
Email, Slack, Discord, generic webhooks, PagerDuty and Opsgenie via webhook. Different severities can go to different places.
Incident timelines
Every incident gets a chronological event log — what tripped, what recovered, which alerts fired, and who acknowledged.
Recurring-pattern detection
Servbyte highlights incidents that share a signature over time so you can fix the class of problem, not the instance.
Sample incident timeline
Incident #4821 · web-03 · high-CPU saturation · resolved in 8 minutes.
- Trigger14:22:10 UTC
CPU on web-03 crossed 85% threshold (rolling 3m).
- Alert14:22:14 UTC
Alert routed to #ops on Slack and PagerDuty (severity: high).
- Note14:23:01 UTC
Anomaly detector: memory usage also diverging from baseline.
- Ack14:24:47 UTC
Acknowledged by @maya.
- Note14:27:52 UTC
AI diagnosis drafted — likely correlated with deploy marker at 14:20:38.
- Recovered14:30:11 UTC
CPU recovered below threshold; incident auto-resolved after 3m stable.
Stop chasing noisy alerts
14-day free trial. Bring your Slack, Discord, or PagerDuty webhook.