Skip to content

Notifications

Notifications e-mail a task's owner and extra recipients when a run succeeds, fails, recovers or is cancelled.

For e-mail to the people who use a flow's output, use subscriptions instead.

How it works

A task notification is sent only when all three switches are on:

Level Switch Default Where
Service notifications.enabled Off MetaMiner Administrator (MMA), Tools > FSS Configuration... > Properties
Environment Email notifications Off Web interface, Admin > Settings > General (Flow Scheduler admins)
Task Enable email notifications for this task Off Notifications tab of the task or chain

When a run ends, the Flow Scheduler stores the e-mail before sending it, then delivers it through your mail server. Stored messages survive a restart. In a high-availability setup, only the leader node sends e-mail.

Before you begin

  • SMTP is configured in MMA under Preferences > Settings > Email: host, port, sender address, and login if your server needs one. The Flow Scheduler reads these settings and picks up changes within notifications.smtp.config.cache.minutes. The e-mail on/off check box on that page does not affect Flow Scheduler e-mail.
  • The MetaMiner Server host can reach the mail server. When secure SMTP is selected in MMA, the mail server must offer STARTTLS; the Flow Scheduler does not fall back to plain text.
  • Task owners have an e-mail address in Tableau, or you enter one as an override.

Steps

  1. In MMA, choose Tools > FSS Configuration..., open Properties, set notifications.enabled to true, click Apply, and restart the Flow Scheduler (Windows service MetaMinerFSS).
  2. In the web interface, open Admin > Settings > General and switch Email notifications on for the environment. This is the environment's switch for task notification e-mail. Turning it off asks "Disable all notifications?".
  3. Open the task (Edit) or chain and go to the Notifications tab.
  4. Select Enable email notifications for this task.
  5. Under When to send, choose one or more:
    • On success - run completed or linked chain succeeded
    • On failure - run failed, lost visibility, or start-fail
    • On recovery - held SERVER_DOWN or recovery-skipped
  6. Set the recipients:
    • The Owner row shows the task owner (or, without one, the creator) and their e-mail, or "no email on file".
    • Owner email override (optional): addresses to send to instead of the owner's.
    • Additional recipients: addresses copied on every notification.
  7. Click Save.
  8. Click Send test e-mail. The result reads "Queued for delivery."

A cancelled run sends a notification whenever the task's notifications are on.

Delivery and retry

Situation What the Flow Scheduler does
Mail server accepts the message Marks it sent
Temporary error (server busy, timeout, connection refused) Tries again after 30 seconds, 1, 5, 15 and 30 minutes, then 1, 2, 6, 12 and 24 hours
Some recipients accepted, others not Retries only the recipients that were not accepted
Permanent rejection Marks the message failed; no retry
The Flow Scheduler restarts before sending Sends after the restart
The same event repeats within notifications.dispatch.throttle.window.minutes Sends one message

To pause all sending without losing messages, set notifications.delivery.enabled to false. Queued messages wait until you set it back.

E-mail content

Setting Effect
notifications.subject.prefix Text in front of every subject
email.app.name Product name in the e-mail header and footer
notifications.fss.external.url Public address of the web interface, used for links in e-mails. Empty = no links. In a high-availability setup, set it on each node, to the load balancer address.

Change these in MMA under Tools > FSS Configuration... > Properties, then restart the Flow Scheduler (email.app.name applies without a restart). See Configuration Reference.

Administrator alerts

Besides task notifications, the Flow Scheduler e-mails administrators about service problems. Like task notifications, alerts use the SMTP settings from MMA.

Alert Sent when Recipients and settings
Tableau back end unreachable An environment's Tableau back end cannot be reached; repeated at most once per env.failure.notify.throttle.sec per environment The support e-mail recipients configured in MMA
Cluster events (high availability) By default: a node becomes leader, a node goes offline, or the leader's MMS is down. The list is ha.alert.email.events. ha.alert.email.to; when blank, the support e-mail recipients configured in MMA

Defaults: Configuration Reference.

Verify

  • The test e-mail arrives.
  • After a run, the message arrives with the subject prefix.

Troubleshooting

No e-mails arrive

Check in order: notifications.enabled is true and the Flow Scheduler was restarted; Email notifications is on for the environment; the task's Notifications tab is enabled with the right events; the SMTP settings in MMA are correct; the owner has an e-mail address. See Troubleshooting.

E-mails arrive late

A temporary SMTP problem puts messages into retry. Fix the mail server; queued messages go out at their next attempt. Check server\logs\flowscheduler.log for SMTP errors.

Next steps