Set up failed event alerts
Snowplow can alert you when a new failed event has occurred. There are two different implementations available to choose from:
-
Failed event alerts run queries directly against the your data warehouse every 10 minutes. This approach keeps warehouse workers continuously active, resulting in higher compute costs compared to classic alerts, but provides faster detection, are richer alert context, and are more configurable.
-
Classic alerts offer cost-effective monitoring by executing queries against our internal telemetry data warehouse tables. This approach minimizes your costs since the computational load runs on our systems.
Feature comparison
| Feature | Classic alerts | Failed event alerts |
|---|---|---|
| Can alert on new failed events | ✅ | ✅ |
| Can send a digest of failed events for a week | ✅ | ✅ |
| Notify via Email | ✅ | ✅ |
| Notify via Slack | ❌ | ✅ |
| Filters | ❌ | ✅ |
| Does not affect pipeline cost | ✅ | ❌ |