Monitors

A monitor is a rule checked each time a run finishes. When the run breaks the rule, Lantfly sends an alert to the bell, email and your alert channels. Use monitors to catch a scraper that starts failing, slows down, returns too few results, loses fields or costs too much.

Where rules live

• On an actor: the actor's Monitoring tab. The rule applies to every run of that actor in your account, including runs from tasks.
• On a task: the Monitoring section at the bottom of the task page. The rule applies only to that task's runs.

Settings → Notifications lists all your monitoring rules in a Monitoring alerts card and links to each one.

Rule types

Rule	Alert when	Setting
Run status	A run fails, times out or is aborted	The statuses to alert on: Failed, Timed out, Aborted
Run takes too long	A run takes longer than the time you set	Longer than (minutes)
Too few results	A succeeded run returns fewer dataset items than expected	Fewer than (items)
Empty fields	In a succeeded run, too many items have one of the fields missing, null or empty	Fields and Empty in more than (%)
Run costs too much	A run is charged more than the amount you set (compute, results and events)	Costs more than (USD)

Empty fields checks the first 1,000 items of the dataset (less when the items are very large; the alert then says how many it checked). Field names can use dots for nested values, such as author.name.

Add a rule

1. Open the actor's Monitoring tab (or the task's Monitoring section).
2. Click Add rule, pick the rule under Alert when, fill in its setting and an optional Name.
3. Click Add.

Each rule can be switched on and off, and deleted (its alert history goes with it). Recent alerts lists the last 50 alerts; click one to open its run. A rule alerts at most once per run.

Where alerts go

Alerts are sent as the actor.monitor notification: to the bell and by email by default, and to the account's alert channels (Slack, Discord, Telegram, Zalo, webhook, email list). Each member chooses in Settings → Notifications. See Notifications.

API

Alert when a task returns fewer than 50 items

curl -s -X POST "https://lantfly.com/api/v3/monitors" \
  -H "Authorization: Bearer $LANTFLY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "task_id": "YOUR_TASK_ID",
    "name": "Daily crawl too small",
    "type": "item_count",
    "params": { "min_items": 50 }
  }'

type	params
run_status	statuses: any of FAILED, TIMED-OUT, ABORTED
duration	max_secs
item_count	min_items
field_empty	fields (list), max_empty_pct
cost	max_usd

Send actor_id (ID or slug) or task_id. Other calls:

Request	What it does
GET /api/v3/monitors	Rules (filter with actor_id, task_id) and their last 50 alerts
PATCH /api/v3/monitors/:monitor	Change name, is_active or params (type and target are fixed)
DELETE /api/v3/monitors/:monitor	Delete the rule and its alert history
GET /api/v3/monitors/events	Alert history, newest first

An account can have up to 50 rules. In an organization, changing rules needs the permission to manage automation.