# Analytics alerts and notifications

> Set alerts for traffic thresholds, anomalies, goals, missing data, downtime, SSL expiry, Web Vitals and usage, delivered by email, Slack, Teams or webhooks.

An alert is a rule that is checked every 5 minutes. When its condition
becomes true the alert is **Triggered** and its channels are notified once.
When the condition stops being true it goes back to **OK** and the channels
get a "Back to normal" message.

All alerts of a workspace are on its **Alerts** page (workspace menu), which
also has the **Channels** tab. A site's own alerts are listed under
**Alerts** in the site menu and in **Site settings → Alerts**. Creating and
editing alerts needs the `content.write` permission.

## Rule types

| Type | Triggers when | Threshold means |
|---|---|---|
| **Metric threshold** | A metric over the window is above, at least, below or at most a value. Metrics: visitors, visits, pageviews, bounce rate, conversion rate (needs a goal) | The value to compare with |
| **Anomaly** | Visitors, visits or pageviews in the window are far from the average of the same window in each of the previous 4 weeks | Not used. Set the sensitivity instead |
| **Goal completed** | The chosen goal is completed at least N times in the window | Completions (default 1) |
| **No data (script missing)** | The site has received no hits for N hours | Hours (default 24) |
| **Usage of plan** | This month's events reach N% of the workspace allowance. It has no site | Percent (default 80) |
| **Uptime down** | An enabled [uptime check](/docs/features/uptime) of the site is down. It resolves when all are up | Not used |
| **SSL or domain expiry** | A certificate or domain seen by the site's uptime checks expires within N days. The site needs at least one uptime check | Days (default 14) |
| **Web Vitals regression** | The 75th percentile of LCP, INP, CLS, FCP or TTFB over the window is past a limit. Needs at least 5 samples in the window | Milliseconds (a score for CLS) |

## Rule settings

| Setting | What it does |
|---|---|
| **Name** | Shown in the list and in every notification. Up to 120 characters |
| **Type** | One of the rule types above |
| **Site** | The site the rule watches. Every type except usage needs one |
| **Metric** | The metric for threshold, anomaly and Web Vitals rules |
| **Condition** | **is above**, **is at least**, **is below** or **is at most**. For an anomaly, above or at least means only spikes, below or at most means only drops, and **Either direction** means both. A Web Vitals rule with no condition uses is above |
| **Threshold** | See the table above |
| **Window** | How far back the rule looks: 15 or 30 minutes, 1, 3, 6 or 12 hours, 1 day or 7 days. Used by threshold, anomaly, goal and Web Vitals rules. The API accepts any `window_minutes` from 5 to 10,080 |
| **Sensitivity** | Anomaly only. **Low** triggers on big changes only, **High** triggers more often, **Medium** is in between. A quiet window (under 5 for both the value and the average) never triggers |
| **Goal** | The goal for a goal rule or a conversion rate threshold |
| **Uptime check** | Optional check id (`upc_…`) for uptime and SSL rules. Leave it empty to watch all of the site's checks |
| **Notify** | The channels that get the notifications |
| **Enabled** | Untick to pause the rule. A paused rule isn't checked |

## Channels

Alerts are delivered through **notification channels**, set up once per
workspace under **Alerts → Channels** and reused by any rule:

| Channel | Plan | Settings |
|---|---|---|
| Email | All | Up to 20 addresses |
| Slack | Pro and up | An incoming webhook URL on `hooks.slack.com` |
| Microsoft Teams | Pro and up | A Workflows webhook URL |
| Discord | Pro and up | A channel webhook URL |
| Webhook | Pro and up | Any public HTTPS URL that accepts JSON, and an optional signing secret. With a secret, requests carry `X-Privatus-Signature: t=<unix time>,v1=<HMAC-SHA256 of "t.body">` |
| Browser push | Pro and up | No settings. Sends to every member who turned on browser notifications on the Channels page |
| PagerDuty | Business | The Events API v2 integration key of a service |
| Opsgenie | Business | An API key and the region (US or EU) |

A channel's type can't be changed after it is created. Each channel has a
**Send test** button (up to 10 tests an hour per member, together with
email report tests).

Every triggered or resolved alert is also sent to the workspace's
[webhooks](/docs/api/webhooks) as `alert.triggered` and `alert.resolved`.
Goal and usage alerts add `goal.completed_threshold` and `usage.threshold`.

Email is limited so a flapping rule can't flood an inbox: one "triggered"
email per alert and address every 30 minutes, 6 emails an hour per alert
and address, and 200 alert emails an hour per workspace. The next email
says how many were held back.

## Noise control

- **Mute** an alert for 1 hour, 1 day or 1 week (for example during a
  planned migration). The API takes any number of `minutes` up to 30 days,
  and `0` unmutes. A muted alert is still checked and still records
  history, and workspace webhooks still get its events. Only its channels
  stay silent.
- **History**: every time the alert triggered, what happened, and when it
  resolved. **Acknowledge** marks the latest incident with your name to
  show the team someone's on it.