We have rebuilt the alerts system with a more intuitive interface and powerful new capabilities. The new experience is available in beta for metric, log and trace based alert types.
Set multiple severity thresholds within one alert instead of creating separate alerts for each severity level. For example, send a warning when CPU usage exceeds 75% and a critical alert when it exceeds 95%, all in one alert. Each threshold can route to different notification channels.
Route notifications dynamically based on alert attributes using expression-based routing policies. Create one alert and let routing policies determine where notifications go based on labels like service name, team, or environment. Perfect for organizations with shared infrastructure where multiple teams need the same metrics routed to their specific channels.
Group alerts by specific fields to combine alerts with matching field values into a single notification, reducing notification volume while keeping related alerts together.
Configure re-notifications to send periodic notifications while an alert condition remains active, giving you control over notification frequency to balance awareness with reducing alert fatigue.
severity_number > "7" or http.status_code >= "400" would compare values alphabetically instead of numerically, leading to incorrect results.IS NULL and IS NOT NULL checks for proper filtering of JSON column types.