Kisu
All articles

BigQuery

BigQuery alerts to Slack

Turn a SQL condition into a scheduled check and send matching results to your Slack channel.

From an order count to a Slack message

Check every 12 or 24 hours

1. Read the data72 ordersIn the last 24 hours
2. Ask a question72 < 100?Yes · condition met
3. Notify the team
Orders need attentionThe order count is below your threshold.
Example Slack message
Example: fewer than 100 orders in the last 24 hours. Each scheduled check asks the same question. This is an illustration, not live data.

A dashboard helps when someone opens it. A scheduled alert helps when a metric needs attention between visits. With Kisu, you can check a BigQuery condition and send a Slack notification when it is true.

Connect your project and Slack channel

Sign in with Google and activate a subscription. In Configuration, add a BigQuery connection with the project that will run and pay for your queries. Sign-in and data access use separate authorization steps.

Your Google account needs permission to create BigQuery jobs in that project and read the relevant datasets. Use BigQuery Job User and Data Viewer permissions where appropriate. Leave the location on Auto-detect unless you need a specific region.

Create a Slack incoming webhook for the channel that should receive alerts. In Configuration, add a Slack destination, paste the webhook URL, and use Send test to check delivery. Treat the webhook URL as a secret.

Write a condition, not a result table

In Alerts, create an alert and select your BigQuery connection. The query must be one SELECT statement that returns one row and one Boolean column. Put the aggregation and comparison directly in SQL.

SELECT COUNT(*) < 100 AS needs_attention
FROM `your-project.analytics.orders`
WHERE created_at >= TIMESTAMP_SUB(
  CURRENT_TIMESTAMP(), INTERVAL 24 HOUR
)

Replace the table and timestamp column with your own. This example asks whether fewer than 100 orders were created in the last 24 hours. That is an example threshold, not a recommended target for every business.

Use Test query before saving. TRUE means the condition is met. FALSE means it is not met. A NULL result or no result rows means no data, not a healthy check. Tests run real queries, and Google bills BigQuery usage separately from Kisu.

Set the schedule and review results

Choose Every 12 hours or Every 24 hours, select your Slack destination under Notify via, and save. These are scheduled checks, not real-time monitoring.

Kisu queues a notification for every successful matching check, including repeated matches while the same incident stays open. In History, review the check result and notification status separately. A matching condition does not prove that Slack received the message.