Alerting Concepts
Stay ahead of issues with Crate’s built-in alerting system. Monitor your traffic in real-time and get notified instantly when anomalies occur, from error spikes to sudden traffic drops.
Supported Metrics
Crate supports a variety of metric types to help you monitor service health:
- Error Count (5xx/4xx): Trigger when the absolute number of errors exceeds a threshold.
- Error Rate (5xx/4xx): Trigger when the percentage of error requests exceeds a threshold.
- Rate Spike (5xx/4xx): Detect sudden jumps in error rates compared to the moving average.
- Traffic Drop: Notifies you if requests fall significantly below your baseline, indicating potential upstream outages.
- Traffic Spike: Detects unusual surges in traffic that might require scaling or mitigation.
Configuring Alert Rules
Alert rules are configured per organization and can be scoped to specific domains or even individual paths.
1. Define the Scope
Choose whether the alert applies to an entire domain or a specific path (e.g., /api/v1/checkout). This allows you to set stricter thresholds for mission-critical endpoints.
2. Set Thresholds & Windows
Configure the Value that triggers the alert and the Time Window (in minutes) over which the metric is evaluated. Longer windows reduce “flapping,” while shorter windows provide faster notifications.
3. Notification Channels
Crate supports multiple destinations for alerts:
- Email: Get detailed reports directly in your inbox.
- Webhooks: Forward alerts to your own internal systems.
Cooldown & Resolution
To prevent notification fatigue, every rule has a Cooldown Period. Once an alert fires, it will not notify you again until the cooldown expires or the issue is resolved.
Crate also sends Resolution Notifications automatically once the metrics return to healthy levels, so you know exactly when the incident is over.