Send a SQL alert straight to Slack.

Before you start

A Slack workspace you can add a destination for, and QueryFlow Pipelines. If you haven't connected Slack yet, do that under Destinations first.

Step by step

1. Add a Slack destination under Destinations if you haven’t already, and send a test message to confirm it’s wired up.
2. Run the query you want to alert on, then click Watch this.
3. Set up the watch: what to watch, a Condition, and Check every.
4. Under Destinations, tick your Slack destination.
5. Click Run preview, then Save.

A worked example

On a Snowflake connection tracking failed payment attempts:

SELECT COUNT(*) AS failed_payments
FROM analytics.payments
WHERE status = 'failed'
  AND attempted_at >= DATEADD(hour, -1, CURRENT_TIMESTAMP());

Set A value on failed_payments, Condition Greater than 20, Check every 1h, and tick your team's Slack channel as the destination. Cross that line and the channel gets a card with the count and the time it fired, instead of someone noticing a revenue dip hours later.

Check it worked

Send a test message when you first add the Slack destination, before relying on it for a real watch. After saving the watch itself, use Check now to fire it once and confirm the message actually lands in the right channel.

Troubleshooting

Slack alerts vs. the Daily Recap

A Slack alert from Watch This fires the moment a condition trips. The Daily Recap is a separate, once-a-day summary of everything that ran and fired overnight. Most teams use both: individual watches for things that need attention now, the recap for the calmer daily read.

Choosing a channel

Point time-sensitive watches, like a payment failure spike, at whatever channel your team actually watches during the day, not an archive channel nobody checks. A quieter, lower-urgency watch can go to a dedicated alerts channel that people glance at rather than get pinged from directly.

What the card includes

The Slack card names the watch, the value or count that triggered it, and the time of the check that fired it, enough context to decide whether it's worth acting on immediately without opening QueryFlow first. For the underlying query and full history, open the watch in the Pipelines sidebar.