Alerts tell you when an app's metrics need attention. Vexo has two kinds:
Every alert is off until you turn it on. You set alerts for each app, in its App Settings, and only the app's owner can change them.
| Alert | Fires when | Default threshold |
|---|---|---|
| Error spike | Your mobile app reports at least this many errors in the last hour. | 500 errors |
| Crash-free sessions drop | The share of crash-free sessions in the last 24 hours falls below this percentage. | 99% |
| D1 retention falloff | D1 retention, over the last 30 days of new users, falls below this percentage. | 20% |
A custom event alert fires when a custom event is sent at least a set number of times in an hour. You can watch as many events as you like, one alert for each event.
Recent custom event alerts lists the custom event alerts that fired in the last 30 days.
Vexo compares yesterday's value of each metric, in UTC, with its daily values over the previous 28 days. An alert fires when yesterday is further from the average than the sensitivity that you set, measured in standard deviations.
| Alert | Fires when the metric | Default sensitivity |
|---|---|---|
| Event volume anomaly | Rises or drops | 3 |
| Active users anomaly | Rises or drops | 3 |
| Crash-free rate anomaly | Drops | 2.5 |
| Error rate anomaly | Rises | 3 |
An anomaly alert needs at least 7 days of data before yesterday, so it can't fire for a new app. Days without data don't count toward the baseline, so an app that stops sending data entirely doesn't trigger an alert. A metric that never changes doesn't trigger one either.
Recent anomalies lists the anomaly alerts that fired in the last 30 days, with the value and its usual average.
Vexo checks alerts every hour, on the hour. After an alert fires, it waits 6 hours before it sends that alert again. When an alert fires, Vexo sends it to:
A webhook receives the alert as JSON:
{
"event": "alert",
"alertType": "ERROR_SPIKE",
"label": "Error spike",
"appId": "...",
"appName": "My App",
"unit": "errors",
"triggeredAt": "2026-11-07T14:00:00Z",
"value": 612,
"threshold": 500
}
A custom event alert also has eventName. An anomaly alert has baseline, the usual daily average, and zScore, the number of standard deviations from it, in place of threshold.