Notifications

Notifications tell you when something in the platform needs attention. Rather than a single
"email me on errors" switch, you build rules: each one subscribes to a specific condition,
scoped to specific things, at a priority you choose, delivered to the people you name.

Two kinds of rule

	Organization rules	My rules
Configured at	Organization → Notifications	Account → My Notifications
Owned by	The organization	You
Applies to	Everyone named as a recipient	You
Who can edit	Users with the Notifications Edit permission	You

Both are evaluated together. If an org rule and one of your own rules both match an event,
you get one notification carrying the highest priority of the two — not two emails.

What you can be notified about

Integrations

Trigger	Fires when
Integration failed	An execution ends in a failed state
Integration succeeded	An execution completes successfully
Integration not run	A scheduled integration hasn't executed within a time window you set

Integration not run is the one that catches silent failures — a feed that stopped being
triggered at all produces no failure to alert on.

Agent Jobs

Trigger	Fires when
Agent job failed	An execution ends in a failed state
Agent job succeeded	An execution completes successfully
Agent job returned no data	The job succeeded but produced no data files

Interface Server

Trigger	Fires when
Interface started	An interface comes up
Interface stopped (error)	An interface stopped because of an error, crash or retry exhaustion
Interface stopped (any)	An interface stopped for any reason, including a user stopping it
Message failures	A threshold of failed message iterations is crossed within a time window
Messages not received	Fewer than a minimum number of messages arrive within a time window

Scripts

Trigger	Fires when
Script execution failed	An organization script execution ends in a failed state

Building a rule

Choose a source — Integration, Agent Job, Interface Server, or Script Execution.
Choose a trigger from that source's list. Some triggers take parameters, such as a
time window and a threshold.
Set the scope — All (every integration / agent job / interface) or Specific
(an explicit list you pick).
Set a priority — Normal, High, or Critical. Priority drives how the notification is
labeled and styled, how it sorts, and how it's grouped for digests and throttling.
Add recipients — either platform users (their contact details are resolved when
the notification is sent) or free-form email addresses for people outside the platform.
Choose channels — in-app and email; SMS is gated on per-user consent

Delivery history

Each notification records, per recipient and per channel, whether it was sent, skipped
(intentionally not delivered — for example, no address could be resolved) or failed (the
channel attempt errored). Check here first when someone reports not receiving an alert.