How do I get alerted when traffic drops or spikes?

Two ways to get alerts

Both send to the notification channels on the notifications page: Webhook, WeCom, DingTalk, Feishu or email.

Create a monitor

1. Open All-platform analyticsOperationsMonitors and create a monitor.
2. Pick what to watch: a built-in measure such as visitors (uv), pageviews (pv), sessions, bounce rate, average duration, error rate or errors, or one saved metric.
3. Pick a grain: hour, day, week or month.
4. Pick a condition (see below) and the channels. If you pick none, every enabled channel is used.
5. Dry run. It calculates without sending and tells you how many times the rule would have fired in the last 30 days. Adjust until that number feels right, then save.

Which condition to use

For traffic drops, a period change on daily visitors is the simplest start, for example down 30% or more. Add a collection gap monitor so a missing script is caught within the hour.

Avoid noisy alerts

• A notification is not repeated for about 60 minutes, and three hits in a row pause the rule.
• Use a day grain for small sites; hourly numbers swing too much.
• A period-change rule stays quiet when the previous period has no data.
• Forecast and anomaly rules show nothing until enough history has built up. That is expected.

Who can set them up

Anyone on the project can see monitors. Creating, pausing, resuming and dry-running need an analyst or above; deleting needs an admin or the owner. Saving notification channels needs an admin or the owner.

When an alert fires, mark the cause on the chart with an annotation.

See also: what to check after a sudden traffic drop and scheduled reports.