# System Feed Ingestion

The System Feed uses two approaches for data ingestion:

**1. Detection Rules** (runs every 30 minutes)

* Custom detection rules you create in the UI
* Uses the pipeline: Source → Filter → Aggregate → Evaluate → Fingerprint → Store
* Executes against your linked integrations (Zuora, Stripe, Salesforce, etc.)

**2. AI-powered Anomaly Detection** (runs every 15 minutes)

* Automatic pattern detection across ingested events
* Uses AI to generate alerts from detected anomalies
* Auto-resolves alerts when issues are fixed

Both approaches feed into the same System Feed. The detection rules give you precise, configurable control over what to monitor, while the AI system catches patterns you might not have explicitly defined.

| Source                   | Schedule     | How It Works                                        |
| ------------------------ | ------------ | --------------------------------------------------- |
| **Detection Rules**      | Every 30 min | Runs your configured rules against integration data |
| **AI Anomaly Detection** | Every 15 min | AI analyzes event patterns and generates alerts     |

So if you create a detection rule for "Failed Payments > 5", that runs via the rules engine. Meanwhile, the AI system might independently detect unusual payment failure patterns even without a specific rule.