Skip to content

Pulses

A Pulse is the atomic unit of monitoring in SilentPulse. Each pulse connects to one data source and checks whether expected assets are reporting.

Think of a pulse like an EKG heartbeat — it periodically checks a data source and verifies that your assets are still sending telemetry.

Pulse "Kafka Windows Events"
├── Connector: Kafka (brokers, topic)
├── Check every: 5 minutes
├── Alert if silent for: 15 minutes
├── Hostname field: hostname
└── Status: Active ✓ (47 assets monitored)

Click Pulses → New Pulse in the sidebar. You need three things:

  1. A name — e.g., “Production Splunk”
  2. Connection details — URL, credentials for your data source
  3. Monitoring mode — Discovery (auto-learn) or Manual (explicit CMDB mapping)

The system connects to your data source and learns which assets are reporting. After a configurable learning phase (1h–7d), monitoring activates automatically.

You select an existing Asset Group from your CMDB. The system immediately starts monitoring those specific assets.

A Pulse is standalone — it monitors one data source independently.

A Pipeline connects 2+ pulses monitoring the same assets through different stages:

Windows Server → Winbeats → Kafka → Splunk → Databricks
Pulse 1 Pulse 2 Pulse 3 Pulse 4

When you have multiple pulses for the same asset group, SilentPulse suggests connecting them into a pipeline. This enables cross-pulse correlation — if an asset appears on Kafka but not on Splunk, you know exactly where the data pipeline broke.

StateDescription
LearningDiscovery mode active, collecting baseline
ActiveMonitoring active, generating alerts
DegradedSome assets not reporting
SilentNo data from source
DisabledManually paused
Terminal window
# Create a pulse with discovery mode
curl -X POST /api/v1/pulses \
-d '{"name":"My Splunk","connector_type":"splunk","connection_config":{"base_url":"...","token":"..."},"discovery_mode":true,"learning_duration":"24h"}'
# List all pulses
curl /api/v1/pulses
# Check if pulses can form a pipeline
curl /api/v1/pulses/{id}/suggest-pipeline