Alert Management & Notifications

Cockpit incorporates a real-time alerting system that monitors physical hypervisors and virtual machines. The alerting engine evaluates metrics against configured thresholds and dispatches notifications to system administrators when anomalies or failures are detected.

---

Alert Evaluation Engine

The alerting subsystem processes telemetry through a three-stage pipeline:

1. Metric Aggregation: Background daemons query real-time physical metrics (such as CPU temperature, processor load, and disk space utilization) and hypervisor system states.
2. Threshold Evaluation: Collected metrics are evaluated against administrative rules (e.g., triggering an alert if CPU utilization exceeds 90% for a continuous duration of 5 minutes).
3. Alert Dispatching: If a rule threshold is breached, the engine generates an alert payload and routes it through configured notification channels (such as SMTP or SNMP).

---

Severity Classifications

Alerts are classified into four severity levels to prioritize system responses:

• Info (Informational): Operational events indicating normal state changes (e.g., successful backup completion or database synchronization).
• Warning: Non-critical events requiring review (e.g., a storage pool reaching 85% capacity or elevated network latency).
• Critical: System failures requiring immediate administrative intervention (e.g., a hypervisor host disconnecting or a virtual machine crash).
• Fatal: Severe infrastructure outages (e.g., an entire storage datastore going offline or a power failure on a hypervisor node).

---

Configuring Notification Channels

Administrators can dispatch alerts to external monitoring platforms or messaging services.

1. SMTP Email Integration

To configure email notifications:

1. Log into the Cockpit console as an administrator.
2. Navigate to Settings > Notifications > Email.
3. Configure the mail relay properties:
   • SMTP Host: The address of your corporate mail server (e.g., smtp.corp.awan.io).
   • Port: Typically 587 (STARTTLS) or 465 (SSL).
   • Authentication: Credentials for the sending mail account.
   • Sender Address: The envelope sender address (e.g., alerts-noreply@corp.awan.io).
4. Define Recipient Routing Rules to direct specific alert severities to appropriate teams (e.g., routing Warning alerts to the system logging queue, and Critical/Fatal alerts to the on-call pager queue).

2. SNMP Traps

To forward alerts to an external Network Management System (NMS) such as Zabbix or PRTG:

1. Navigate to Settings > Notifications > SNMP.
2. Configure the SNMP target properties:
   • Target Host: The IP address or hostname of the monitoring collector.
   • Port: Typically 162 for SNMP traps.
   • Community String: The credential payload (e.g., a custom read-only string).
   • SNMP Version: Select v2c or v3 (v3 is recommended as it supports cryptographic authentication and privacy).