Alerts
Get notified when a measure on one of your question types moves past the level you set, so you can react before a regression becomes a trend.
Overview
Alerts tell you when a rule finds a measure past the level you set. Rules are checked after each capture, once the question type has enough capture cycles to judge (see When a rule is first checked). Instead of manually checking dashboards, you set your levels once and the platform surfaces the events that matter.
Alerts are built on two pieces:
- Alert rules — the conditions you define (see Alert Rules)
- Alert history — the record of every time a rule fired, shown on this page
The Alerts page
Open Alerts from the sidebar. You'll see:
- Unread count at the top, matching the badge in the sidebar
- Alert feed — a chronological list of every triggered alert
- Filters — by rule, by severity (info / warning / critical), by time range, or by read/unread status
- Mark as read — click any alert to mark it read; the sidebar badge updates
Each alert card shows:
| Field | Meaning |
|---|---|
| Rule name | Which rule fired |
| Severity | Info (FYI), Warning (needs attention), Critical (act now) |
| Metric & threshold | The value that crossed the threshold |
| Triggered at | Timestamp of the event |
| Context link | Deep-link into the snapshot, prompt, or analytics page that caused it |
Notification channels
When a rule fires, alerts can be delivered to:
- In-app feed (always on) — shows up on this page and in the header bell icon
- Email — one alert email per person covering the project's new alerts, sent to the recipients each rule names (organisation owners and admins unless you choose otherwise), if email notifications are enabled for your plan
- Webhook — fire-and-forget HTTP POST to any URL, with HMAC-SHA256 signing. See Webhooks
- Slack — via the Slack connector. See Connectors
Plan gates: Email notifications are available on Starter and above. Webhooks are available on Pro-SME and above. Slack (via the Slack connector) is Pro-SME and above.
Best practices
- Start with a few high-signal rules. Switch on the rules for the question types you act on first, and add more as you learn what matters.
- Acknowledge alerts promptly. Unread counts pile up and dilute the value of the system. Triage at least weekly.
- Route alerts to a shared channel. Email to one person is a single point of failure — use webhooks or Slack for team visibility.
Related docs
- Alert Rules — how to configure the conditions that trigger alerts
- Webhooks — deliver alerts to any HTTP endpoint
- Connectors — Slack, Zapier, and other outbound integrations