When managing production servers, clusters, and microservices, monitoring dashboards manually introduces alert fatigue. When node CPU utilization spikes or memory pressure drifts above critical thresholds, the platform must evaluate incoming telemetry streams and alert on-call engineers automatically.
Omnismith evaluates event-driven automation rules directly against incoming telemetry packets without blocking primary ingestion.
Example Telegram action configuration for a server CPU breach:
{
"type": "telegram",
"config": {
"channel_id": "019fb9ad-c8ae-7026-bc2d-6cffe1d73e27",
"chat_id": "@omnismith_ops",
"message_template": "🚨 HIGH CPU UTILIZATION: {automation.name}\nHost: {values.hostname} ({entity.id})\nCPU Usage: {previous.cpu_usage}% → {values.cpu_usage}%\nTime: {timestamp}"
}
}Key operational capabilities:
• Configurable cooldown periods (e.g.
cooldownSeconds: 300) throttle downstream alerts to eliminate notification storms during sustained spikes.• Dynamic token resolution (
{previous.cpu_usage}% → {values.cpu_usage}%) displays the exact directional delta in the notification body.• Decoupled background event processing ensures zero impact on metric ingestion throughput.
Step-by-step automation guide:
https://omnismith.io/blog/configuring-event-driven-automations-and-telegram-notifications
#devops #infrastructure #monitoring #automation #omnismith
