Alerts
Alerting Overview
Configure real-time threshold rules that notify your Slack, Discord, PagerDuty, or custom backend when production issues occur.
How Real-Time Alerts Work
1
Log Stream Evaluation
Every ingested log is evaluated in NATS JetStream against your active alert rules.
2
Redis Sliding Window
Counts matches in time buckets. Once threshold count is reached, triggers notification.
3
Signed Webhook Dispatch
Sends HMAC-SHA256 signed JSON payload with matched logs and sample error context.
Supported Condition Operators
| Operator | Behavior | Example |
|---|---|---|
| equals | Exact match (case-insensitive) | Type equals “error” |
| not_equals | Negative exact match | Environment not_equals “development” |
| contains | Substring match | Message contains “timeout” |
| not_contains | Substring exclusion | Message not_contains “health-check” |
| greater_than | Importance ranking or numeric threshold | Importance greater_than “medium” |
Alert Cooldowns & Quotas
To prevent alert fatigue and spam during sustained outages, each triggered alert enters a cooldown lock (e.g. 5m, 15m, 1h). During cooldown, further webhooks for that rule are suppressed.