Scalingo Metric Alert Notification Frequency and Duration Constraints

Metric Alerts and Notifiers

Scalingo allows the configuration of threshold-based alerts on per-container runtime metrics, including CPU, RAM, swap, and network traffic. These alerts trigger notifications via email, Slack, or webhooks when a metric exceeds a defined numeric value.

Notification Noise Management

While horizontal autoscaling can mitigate transient spikes by adjusting container counts, human-facing alerts are still susceptible to "flapping" if a metric oscillates around the threshold. To prevent notification fatigue, it is necessary to understand the evaluation logic regarding the persistence of a breach.

Current documentation details the comparison operators and threshold values, but there is uncertainty regarding the temporal constraints of the alert state. Specifically, it is unclear if the platform supports a sustained-breach duration (e.g., firing only after a metric remains above the threshold for X minutes) or a cooldown period to prevent repeated notifications for a single ongoing event.

• Does Scalingo support a minimum duration for a breach before a notification is dispatched?
• Is there a configurable re-notification interval or cooldown period while a metric remains in a firing state?