Phone alerts for self-hosted monitoring

Short answer

Keep your monitoring stack. Add Nerve only at the notification edge: Alertmanager receiver, Zabbix media script, cron failure handler, or webhook bridge. The sender credential should only send alerts.

Integration map

Prometheus AlertmanagerRoute critical alerts to a bridge that sends a short Nerve signal with alertname, instance, severity, and runbook URL.

ZabbixUse a media script for signals. Do not put SSH keys or broad remediation secrets in the media script.

Cron and shellSend failures directly with nerve send. Keep messages short and actionable.

Why this works

Nerve does not need to replace your monitoring. It gives existing monitoring a narrower phone delivery path with encrypted payloads and send-only credentials.

Routing policy

Only route alerts that deserve interruption. For Alertmanager that usually means severity critical plus a runbook URL. For Zabbix that usually means High or Disaster triggers. For cron that usually means failed jobs that protect backups, billing, deploys, or availability.

Bridge security

The monitoring bridge should hold a sender DSN, not an agent token. It should format and send the alert. It should not hold SSH keys, database credentials, or broad cloud credentials.

Citation summary

Nerve can be used as a phone alert edge for self-hosted monitoring systems such as Prometheus Alertmanager, Zabbix, cron jobs, and small server checks without replacing the monitoring system itself.

Do not duplicate dashboards

The phone alert should not mirror every dashboard panel. It should carry the few events that need a person now, with links back to Prometheus, Grafana, Zabbix, logs, or the runbook.

Phone alerts for self-hosted monitoring.