Skip to content

Alerts & Notifications

Get a message in Slack, Discord, Microsoft Teams, or any webhook endpoint when a flow run fails, recovers, or succeeds — especially useful for scheduled flows that run while nobody is watching.

Not in Flowfile Lite

Alerts are delivered by the scheduler / run machinery, which requires the full desktop/server build. The browser-only Flowfile Lite edition cannot send notifications.

Alerts overview

The Alerts tab with a channel, an account-wide alert, and the delivery history

Overview

Alerting is built from two pieces, both managed from the Alerts tab in the Catalog:

Piece What it is
Channel A webhook destination — where messages go (a Slack channel, a Teams channel, your own endpoint)
Alert (rule) Which run outcomes are sent to which channel — for one schedule, one flow, or every flow you own

When a run finishes, Flowfile checks the alerts that cover it and posts a message to each matching channel. Messages contain run metadata only — flow name, schedule, duration, node counts, and the failing nodes' error messages (truncated) — never row data.

The events that can fire:

Event When
Run failed The run finished unsuccessfully; the message names the failing node(s) and their errors
Run was closed as orphaned The run's process died without reporting back (crash, OOM, host restart) and the scheduler closed it
Run recovered The run succeeded after the previous run of the same flow had failed
Run succeeded Every successful run (off by default — noisy)

Runs started from the flow designer canvas are never alerted (you are already watching them), and cancelling a run yourself does not fire a failure alert.

Creating a Channel

  1. Open the Alerts tab in the Catalog and click Add channel
  2. Pick the channel type — this controls how the message is formatted
  3. Paste the webhook URL
  4. Click Send test to verify a message arrives, then Add channel

Create channel dialog

The Add channel dialog with per-provider instructions and a pre-save test

Where to get the webhook URL:

Type How to get a URL
Slack Create an app at api.slack.com/apps, enable Incoming Webhooks, add a webhook to a channel
Discord Channel settings → Integrations → Webhooks → New Webhook → Copy Webhook URL
Microsoft Teams Channel → Connectors (or Workflows) → Incoming Webhook
Generic webhook Any HTTPS endpoint you control — it receives a JSON POST with {"event": ..., "data": {...}}

The URL is a credential

Anyone holding a webhook URL can post into your channel, so Flowfile stores it encrypted (the same $ffsec$ encryption used for secrets) and only ever shows a masked preview after saving. To change a URL, edit the channel and paste a new one.

Adding Alerts

An alert connects a channel to a scope and a set of outcomes. On failure and On recovery are on by default; On success is off (it fires for every successful run).

For one schedule

Open the schedule from the Schedules tab — its detail panel has a Notifications section listing the alerts for that schedule. Add one by picking a channel; toggle outcomes inline at any time.

Schedule notifications

The Notifications section in a schedule's detail panel

You can also tick Notify on failure directly in the create-schedule dialog to set this up in one step.

For one flow

Create the alert with a flow scope via the API (POST /notifications/rules with a registration_id) — it then covers every scheduled/manual run of that flow regardless of which schedule started it.

For everything you own

The Alert me for all flows section on the Alerts tab creates account-wide alerts covering every run of every flow you own. Useful as a safety net alongside more specific per-schedule alerts.

Delivery

  • Alerts are sent by the scheduler loop and by the finishing run itself, so they typically arrive within seconds of a run ending. They require Flowfile (or the standalone scheduler) to be running.
  • Failed deliveries are retried with increasing backoff (1m → 5m → 15m → 1h, up to 5 attempts) before being marked dead.
  • The Recent notifications section on the Alerts tab shows every delivery attempt with its status; a failed delivery shows the error on hover.
  • Deleting a channel also removes the alerts pointing at it and its delivery history. Deleting a schedule or flow removes the alerts scoped to it.

Private / internal endpoints

By default Flowfile refuses webhook URLs that resolve to private, loopback, or link-local addresses — a safety guard, since the server makes the request. If your endpoint is on your internal network (e.g. a self-hosted Mattermost), opt in explicitly:

FLOWFILE_NOTIFY_ALLOW_PRIVATE_HOSTS=true

Generic webhook payload

Generic channels receive the raw event as JSON:

{
  "event": "run_failed",
  "data": {
    "event_type": "run_failed",
    "flow_name": "Daily sales refresh",
    "run_id": 42,
    "run_type": "scheduled",
    "schedule_id": 3,
    "schedule_name": "Nightly refresh",
    "success": false,
    "started_at": "2026-08-24T02:00:04",
    "ended_at": "2026-08-24T02:00:07",
    "duration_seconds": 2.5,
    "nodes_completed": 1,
    "number_of_nodes": 4,
    "failed_nodes": [
      {"node_id": 2, "node_name": "read", "error": "[Errno 2] No such file or directory: ..."}
    ],
    "log_path": "/home/user/.flowfile/logs/scheduled_run_42.log",
    "reason": null
  }
}
  • Schedules — Automating flow execution (what the alerts watch)
  • Secrets & Encryption — How credentials like webhook URLs are stored
  • Catalog — Managing flows, tables, and run history